UMD CS News

Graphics Visualization and VR AR

Technica Brings Students Together for Weekend of Coding

Samuel Malede Zewdu — Wed, 21 Oct 2026 14:02:48 +0000

The annual hackathon brought students together to develop technology projects, explore new skills and connect with industry mentors.

October 21, 2026

Two participants collaborate on a project during Technica 2026 at the College Park Marriott Hotel in College Park, Maryland.

Students gathered at the College Park Marriott Hotel and Conference Center on Oct. 17-18 for Technica 2026, a student-organized hackathon that brought together participants with different levels of programming experience for a weekend of coding, workshops and collaboration.

Now in its 12th year, the University of Maryland event welcomed over 1,000 in-person and virtual participants under the theme "Dare to Dive." Technica is the world’s largest hackathon celebrating underrepresented genders, giving beginners and experienced programmers a space to build projects and learn from one another. Mentors, volunteers and industry representatives offered technical guidance as teams worked toward their final demonstrations.

The weekend concluded with awards recognizing projects that explored a range of technology applications. Winning entries included tools that addressed digital accessibility, personal health management, online privacy and financial literacy, as well as educational games focused on environmental issues. Additional awards recognized beginner projects and entries developed for sponsor challenges.

For Maelle Nghonda, a sophomore graphic design student at Montgomery College, the hackathon offered a chance to apply her interest in programming outside the classroom. Nghonda began experimenting with code in middle school after being introduced to Scratch, a visual programming platform.

"I liked the idea of seeing my creations come to life," Nghonda said. "And I always took at least one programming class from middle school until college."

At Technica, she planned to develop a social networking platform for college students that would help users find study partners, exchange notes and locate textbooks. The concept also included direct communication between students and faculty and a customizable, AI-powered study companion.

Nghonda said she wanted to see how projects take shape in a competitive environment and learn from students at other institutions.

"It's been easy to get along with other students and students from all over Maryland and D.C.," she said.

Darian Vadakkoot, a junior at Westfield High School in Fairfax County, Virginia, brought experience in artificial intelligence, web development and Python to Technica. His team was still considering project ideas when the event began, but he said they wanted to develop something that could address a practical need.

"We hope to work on something useful that people might really need right now, and there are many things we can address,” Vadakkoot said. “This is the right place to make that happen.”

While building a project was part of his goal, he also wanted to expand his knowledge of programming languages and meet others interested in technology.

Participants could seek assistance from mentors and volunteers throughout the weekend, including guidance on developing project ideas and addressing technical challenges. According to Claire Chen, Technica's co-director and a UMD computer science senior minoring in Asian American studies, questions about artificial intelligence have become more common as students explore how to incorporate the technology into their work.

"I've seen a lot more questions in Technica and also other hackathons that I've attended as a hacker regarding how to use AI models and how to use AI to improve workflows," Chen said.

Chen said she views AI as a tool to help participants organize ideas, develop projects and prepare demonstrations, rather than replace the work involved in a hackathon.

Her perspective is also shaped by her experience as a former participant. Chen first attended Technica in high school and said the connections she made there influenced her decision to study computer science at UMD.

"After my early experience at Technica, I realized that I found a safe community and space within the tech field," Chen said.

Having returned to support participants, Chen said the event can influence how students approach technology long after the competition ends.

"One thing that I think people really take with them is the impact and inspiration to pursue a field in technology," she said.

—Story by Samuel Malede Zewdu, CS Communications

Technica 2026 winners

Fin-tastic Design: Best UI/UX Hack — LURE

An ocean-themed interactive experience that teaches users to identify deceptive website designs, known as "dark patterns," through a simulated free trial and an opportunity to redesign misleading interfaces.

A Sea for All: Best Accessibility Solution — Spoken.ai

An application that uses an AI voice agent to contact local service providers, negotiate prices and compare offers, allowing users to request services through photos and text rather than phone calls.

Deep Sea Gamechanger: Best Gamification Hack — Clawcean

An AI-powered claw machine game that teaches students about microplastics and ocean pollution through interactive activities focused on plastic waste and its environmental effects.

Calming Currents: Best Active-Wellness/Health Hack — Nari

An AI-powered menstrual wellness application that helps users track symptoms, identify patterns that may require medical attention and prepare summaries for health care appointments.

Dive into Equity: Best Hack for Social Good — Earth Vanguard

A story-driven space shooter featuring different playable characters and voice-controlled gameplay, including an underwater sequence addressing ocean pollution, with accessibility incorporated into its design.

The Little Mermaid: Best Beginner Hack (Middle and High School) — Ecosphere

An interactive website combining AI-powered questionnaires, ocean visualizations and a simulation game to demonstrate how everyday decisions affect the environment.

Uncharted Waters: Best Beginner Hack (College) — CarePulse

A chronic illness management platform that allows patients to track symptoms, medications and appointments while using AI to organize information for patients and health care providers.

Best Shell-ucation Hack: Best Education Hack — ASL Arena

An interactive American Sign Language game that uses real-time hand tracking to allow players to compete against one another.

Rising Tide: Best Startup Track Hack — ClearSpace

A privacy-focused hidden camera detection tool that uses patterned light flashes and changes in Wi-Fi traffic to identify potential cameras without accessing or storing private user data.

'Small Startup with a Big Vision'

Samuel Malede Zewdu — Wed, 21 Oct 2026 13:17:55 +0000

How Phil Crosby (B.S. ’06, computer science; B.A. ’06, economics) made his lifelong tech startup dream a reality as co-founder of the mobile advertising platform Liftoff.

October 21, 2026

Phil Crosby. Photo courtesy of same.

In June 2026, California tech company Liftoff went public on the Nasdaq Stock Market. For the company, it was a milestone 14 years in the making. For Phil Crosby (B.S. ’06, computer science; B.A. ’06, economics), the company’s co-founder, the adventure started way before that.

“In middle school and in high school, I always knew I wanted to do a startup, so starting the company came naturally from my childhood and my early work experience at the University of Maryland. I just couldn't imagine life without it,” Crosby said. “What I did not imagine was that it would come this far, that this startup we built would grow to be hundreds of people across the world, servicing so many customers that we could go public, and public investors would care about us. It's awesome, but surreal.”

Launched in 2012 by Crosby and co-founders Mark Ellis and Harry Robertson as what Ellis called a “small startup with a big vision,” Liftoff is now a multi-million-dollar global company, making a game-changing impact in the world of mobile advertising with its artificial intelligence-enabled growth acceleration and marketing platform Cortex. The platform helps app developers target new users and optimize their engagement, growth and revenue.

“What we built is a system that uses machine learning and AI to predict what apps someone might be interested in, and then we can show advertisements when they're using other apps, and tell them, ‘Hey, did you know about this app? You should try it—click here to install,’” Crosby explained. “We’ve been able to reliably build on the foundation we started with, and as the performance advantages of our system have compounded, we’ve been able to leave most competitors behind. I'm amazed and thankful for what we’ve accomplished.”

Self-taught and always curious

Growing up in Montgomery County, Maryland, Crosby got his first computer—a 386 PC—in elementary school. Always curious, he taught himself to program and build websites by middle school.

“I made a Calvin and Hobbes fan website, a Street Fighter website, and eventually, I had a graphic design website, and people started contacting me to make logos and websites for them. They were actually paying me, a middle schooler, to do this stuff,” he recalled. “I realized then that I could make money from using computers and building these skills. It was clear even in middle school: I'm going to study computer science, for sure.”

Even then, Crosby was intrigued and impressed by the power of technology.

“In high school, everybody I knew was using Microsoft and their products for work, instant messaging, and playing games,” he said. “I thought, okay, cool, that's how you use technology to affect everybody. It changed the world, and I wanted to do that.”

In the early 2000s, as a computer science and economics dual-degree student at the University of Maryland, Crosby dug deeper into the challenges of software engineering, design and data visualization.

“The depth of the education I got at UMD was very helpful, and I think the most important class I took was human-computer interaction. It was my first time learning from somebody about how software should make you feel, how a human should understand it – not the visual design, but the interaction design – how the workflow is supported by the software. I also learned data visualization—how to take a large dataset and present it so you can derive insights from it,” Crosby explained.

Crosby and his roommate Michael Quinn (B.S. ’06, computer science) were in the data visualization class together.

“We’d stay up all night doing the homework—and now, he’s one of the early employees at Liftoff, can you believe it? The community at UMD was really important for me,” Crosby reflected.

Impactful undergraduate internships at Microsoft and IBM reinforced Crosby’s plans.

“Those internships were pivotal for me to see my career, what it could be and what it takes to be a good software engineer,” Crosby said. “I saw what big company life is like, which helped reinforce that I definitely wanted to do a startup; something small where I can have more influence and impact.”

On to Silicon Valley

Anxious to engage with more tech innovation, Crosby headed to California after graduation.

“I moved to Silicon Valley with just a suitcase and no money, and a tenuous agreement to do some consulting work with a small lab, but, oh, it was glorious,” Crosby said. “All these internet startups were blowing up, and I was in this little office doing contract coding work right next to Facebook's headquarters. It was a really cool time.”

In 2008, Crosby joined Ooyala, an online video technology innovator in the heart of Silicon Valley that supported big media companies like NBC and ESPN. There, he leveraged his user experience and data science skill set and expanded his software engineering skills to develop new products, including the frontend system for a Flash player that was used by millions of people.

But he hadn’t forgotten his plan to launch a startup. In 2011, a company hackathon became the catalyst.

“Harry, Mark and I worked at Ooyala together, and that year, the company held this hackathon where the challenge is to build a new idea that's related to the company. We spent two nights, all night, coding in the office—and we won!” Crosby recalled. “Afterward, it was clear. I said, ‘Guys, we have to start a company together!’”

At that time, in the tech world, countless apps were launching, and app stores were booming. But Crosby and his colleagues could see something was missing: a way to connect these apps to the people who would want to use them. That problem launched Liftoff.

“Apps were being developed like crazy back then, but we noticed that even if you had a marketing budget, that didn't mean you could reliably get an audience for your app. For larger companies like Amazon or Facebook, and the mid-sized companies in between, transitioning their products from the web to mobile apps was really difficult,” Crosby explained. “If you make a hit app, how will people come? That’s the problem we wanted to solve.”

So, Crosby and his co-founders tapped into their combined experience—from building mobile apps and video delivery systems to analytics and monetization—to find the answer. And through years of problem-solving research, engineering and innovation, the Liftoff team created one of the world’s biggest independent mobile app growth platforms, supporting more than 167,000 apps in 200 countries.

And now, the ‘small startup with a big vision’ isn’t so small anymore.

“We started with the three of us, and Liftoff was fewer than 20 people for the first three years or so, and those days were so fun,” Crosby said. “Today, we are over 650 people, and we have offices around the globe. It’s still hard to believe sometimes.”

For Crosby, the startup experience has gone far beyond anything his middle-school self could have hoped for—and the adventure continues.

“I think it's one of life's greatest adventures to be able to start a company, because you have the autonomy to make choices, and the success or failure is on you. And it's a very social and connecting activity. You hire the team that you want to work with and spend all your time with them trying to overcome difficulties and adversity, and that bonds you—like at university, the experiences really bond you together,” Crosby said. “So I think it's one of the great things that someone can do with their life. I've really enjoyed it, and I think Liftoff will keep on rocking from here.”

—Story by CMNS

Smarter Estate Settlements

Samuel Malede Zewdu — Mon, 19 Oct 2026 19:08:41 +0000

For Jacqueline Deprey (B.S. ’20, computer science; B.S. ’20, operations management and business analytics), a personal loss and a passion for problem-solving inspired Perpetua, a digital toolkit designed to simplify estate settlement.

October 19, 2026

Jacqueline Deprey. Photo courtesy of same.

From her first coding experience in high school, Jacqueline Deprey (B.S. ’20, computer science; B.S. ’20, operations management and business analytics) has always been drawn to the problem-solving power of computing.

“I don't know that I ever loved coding, but once I understood the impact it could make on other people, I just developed a passion for it that I couldn't give up,” she said. “And that's what motivated me to major in computer science at Maryland.”

This year, leveraging the software engineering and business expertise she built over the past decade at the University of Maryland and beyond, Deprey turned her passion for computing into a business.

She and Shruti Das (B.S. ’21, computer engineering) co-founded the tech company Perpetua, which takes on one of life’s daunting—and often heartbreaking—challenges: settling the estate of a lost loved one. The company’s digital toolkit is designed to answer the tough questions families face when a loved one passes away.

“How can we help families capture both the financial and emotional value of the things their loved ones leave behind? We tried to use technology to basically create a playbook,” Deprey explained. “Perpetua helps families and advisors track down and gain access to the known and unknown online accounts of those who have passed, whether that means bank accounts, medical records, insurance policies, social media, a pension plan or even a life insurance policy you might not know existed. It finds the accounts, subscriptions and assets tied to a life and shows you exactly what to do next.”

It’s not the business Deprey envisioned launching as a student, but after her family experienced this pain firsthand, she couldn’t turn away.

“This problem is just calling me,” she said. “I can't think of anything more impactful to be doing with my time.”

Charting a course

For Deprey, UMD was a part of her life long before she became a college student.

“Both of my parents were Terps, and I remember, even as a small child, knowing the Maryland fight song just as well as I knew the Pledge of Allegiance,” she said.

From day one at UMD, Deprey knew she wanted to pursue computer science and business. Participating in the Gemstone and QUEST Honors programs helped her decide on a path for the future.

“I wasn’t sure if academia or industry was going to be my path after college. I found that I really loved being in Gemstone, getting to do hands-on research that was making an impact. Working on an interdisciplinary team allowed me to learn from students who were different from myself, and we were able to build an assistive device to help children with learning disabilities, which was very cool,” she said. “Then through QUEST I was able to work hands-on with different companies and do similar interdisciplinary teamwork, but through an industry lens, which helped shaped the kinds of problems I wanted to solve after I graduated.”

At Maryland, Deprey stayed busy and involved. Active as a tutor and teaching assistant in what is now the Computing Catalyst, she was also a member of the Omicron Delta Kappa honor society, a campus tour guide and even president of the table tennis team. A junior-year internship at Dropbox landed Deprey a full-time job after graduation, working first as a software engineer and later as a senior product manager.

An academic BOGO—and a search for solutions

In 2024, with ambitious plans for her future, Deprey headed to Columbia University to earn her MBA and a master’s degree in engineering.

“Because I had stopped hands-on coding as a software engineer, I wanted to sharpen the technical skills I had first built at Maryland. And when I looked at people who had the influence that I wanted later on in my career, I noticed many of them had those 3 letters next to their name: MBA,” Deprey explained. “I jokingly describe it as a BOGO. I’d get to understand how AI works under the hood and increase my business acumen all at the same time.”

While at Columbia, a sudden family loss changed everything. When a close relative passed away, Deprey saw the devastating impact it had on her entire family and the challenges they faced trying to settle the estate. Still dedicated to tackling tough problems, Deprey knew she had to do something.

“Seeing how stressful this estate process is, I kept thinking there has to be a better way,” Deprey said. “At Dropbox I built systems that helped organizations manage employee accounts after someone left a company—what should be preserved and what should remain private—and I couldn't help but think, if there's a solution out there that helps with departed employees, why isn't there a solution that's helping families when they have departed loved ones?”

Months later, after countless conversations with estate lawyers, legal experts, tech companies and a host of people who had struggled through an estate settlement themselves, Deprey and Das created Perpetua, an online toolkit designed to make the process easier.

“Perpetua’s direct-to-consumer offering is twofold,” Deprey explained. “For those who want to plan ahead, there is a free guided playbook that takes about 20 minutes to complete. Think of it as a practical companion to your legal documents that you can share with someone you trust. For those who have recently lost someone, Perpetua helps families track down and find the accounts tied to their loved one’s online presence and get access to them after the person has passed.”

It’s problem-solving software Deprey wishes her own family could have had when they needed it most. Though building Perpetua has been a challenging and sometimes emotional journey, she’s grateful for the invaluable learning experiences at UMD that helped make her problem-solving platform a reality.

“To be able to help others is really what drives me, and I'm so fortunate for the foundation that Maryland gave me that helps me do that in such a tangible way,” Deprey said. “Whether it was the technical skills needed to code products like this or the business insights needed to bring Perpetua to market, knowing I could ease someone's burden, even just a little bit, in their time of grief means the world."
—Story by CMNS

William Regli Receives Fulbright Specialist Award

Samuel Malede Zewdu — Mon, 19 Oct 2026 16:44:04 +0000

He will collaborate with the University of Melbourne on artificial intelligence risks, national security and policy development.

October 19, 2026

As artificial intelligence becomes more widely used in national security, government operations and other critical areas, questions about its safety and reliability are drawing increased attention. University of Maryland computer science Professor William Regli (Ph.D. '95, computer science) will examine these challenges through an international collaboration in Australia.

Regli has received a Fulbright Specialist Program award from the U.S. Department of State and the Fulbright Foreign Scholarship Board. Through the program, he will conduct educational and training activities in computer science and information technology at the University of Melbourne, focusing on risks associated with AI and autonomous systems.

During his appointment, Regli will work with Melbourne National Security and Defence (MNSD), a new function to strengthen the University of Melbourne’s capability to partner on research and education supporting Australia’s national security. His activities will support research into the risks of using AI in mission-critical environments, where system failures or unintended outcomes could have significant consequences.

“My specific activity will be supporting their research on AI and the nature of AI and Autonomy risks, especially in the context of mission-critical applications,” Regli said. “I hope to contribute to the development of policy frameworks that decision-makers can deploy to analyze AI use cases and emerging applications.”

MNSD Executive Director Professor Michael Webb said the partnership brings together vital expertise, research and leadership in areas of increasing global significance.

“Professor Regli is highly regarded as an academic leader in the fields of AI and national security – both rapidly evolving challenges and strategic priorities for our nation and our University as we support sovereign defense capabilities with openness and integrity,” Professor Webb said.

The collaboration aligns with Regli's research on AI safety, intelligent systems and the relationship between technology and human decision-making. Regli’s work examines how organizations can assess the reliability of AI technologies while accounting for the environments and circumstances in which they operate.

“Given that AI is everything, everywhere, all at once, we are being overwhelmed with AI's application and insertion into all manner of human activity,” Regli said.

In recent years, Regli has approached AI safety from a systems-level perspective, examining technical performance alongside human and social factors. Rather than evaluating AI in isolation, his work considers how risks emerge in specific applications and the broader systems in which they are deployed.

His research draws on his academic and federal science policy experience. From 2014 to 2017, Regli served in leadership positions at the Defense Advanced Research Projects Agency (DARPA), including deputy director and acting director of its Defense Sciences Office.

He later served as founding executive director of UMD's Applied Research Laboratory for Intelligence and Security from 2018 to 2023. From 2023 to 2024, he worked as a senior adviser for AI risk in the White House Office of Science and Technology Policy.

In addition to his government service, Regli has conducted research spanning artificial intelligence, engineering design, manufacturing and computational modeling. Since rejoining the department in 2018, he has mentored several doctoral students working in areas as diverse as transfer learning in robotic systems, orchestration of computational workflows, and multi-agent coordination.

The Fulbright Specialist Program connects U.S. faculty members and professionals with institutions abroad for short-term projects involving education, research and professional development. More than 400 Americans participate annually, with recipients selected based on their professional experience and potential to contribute to international institutional cooperation.

The program is part of the broader Fulbright Program, established in 1946, which supports international educational and cultural exchange in more than 160 countries.

“As a Fulbright Scholar, it is an honor to be supporting the State Department's international scientific diplomatic efforts and represent the University of Maryland in this way,” Regli said. “I hope that these activities open opportunities for international collaboration for our students and faculty, as well as for those at the University of Melbourne.”

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Helping Robots Make Sense of the Physical World

Samuel Malede Zewdu — Tue, 13 Oct 2026 17:03:14 +0000

October 13, 2026

A glass wall can look like an open path to a robot. Reflections and transparency can fool its sensors, leaving it with a faulty map—and a route that could end in a collision.

University of Maryland computer science doctoral student Naitri Rajyaguru and collaborators are helping robots recognize those barriers using information ordinary cameras miss: the polarization of light.

Their work earned two honors at workshops associated with the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), held Sept. 27–Oct. 1 in Pittsburgh. In all, UMD’s Perception and Robotics Group received five honors across five workshops at the flagship robotics conference.

The honors recognized research led by Rajyaguru and fellow computer science doctoral students Zhi “Leo” Wang and Amir-Hossein Shahidzadeh. Their three projects address how robots perceive their surroundings, learn from people and track an action’s progress.

The students are advised by the group’s co-directors, Yiannis Aloimonos, a professor of computer science, and Cornelia Fermüller, a research scientist in the University of Maryland Institute for Advanced Computer Studies (UMIACS). Aloimonos also holds an appointment in UMIACS, which provides administrative and technical support for the group.

“Our group has been at the forefront of research connecting language, perception and robotic action for many years,” said Aloimonos. “What makes this latest recognition especially meaningful is that it highlights the emerging researchers in our lab. These students are bringing fresh ideas to longstanding challenges and helping shape the next generation of intelligent robots.”

Rajyaguru led “PolarDepth: Polarization-Guided Monocular Depth for Visual Odometry,” which earned the Best Poster Award at the Insect-scale Autonomy Workshop and Best Poster Runner-Up recognition at the Bridging Perspectives in Navigation: Lessons Learned and Future Directions Workshop.

PolarDepth improves distance estimates and movement tracking in spaces lined with glass. Rajyaguru and co-authors explain that polarization—the orientation of light’s oscillations—provides clues to surface geometry that conventional images can miss.

Using a single polarization camera, the system combines color imagery with polarization cues. It learns which is more reliable in different parts of an image, drawing on polarization around reflective surfaces and color imagery elsewhere.

Tests on real and synthetic scenes showed better depth estimates and movement tracking in glass-walled corridors than approaches using color images alone. That could help robots navigate buildings filled with glass doors, walls and partitions.

The Perception and Robotics Group’s other recognized projects address learning and interacting with objects.

Wang led “HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos,” which received the Best Paper Award at the WORLDS: World Models and Spatial Intelligence for Physical AI Workshop and the Best Demo Paper Award at the Building Scalable Infrastructure for Robot Learning: From Data Scaling to Real-World Deployment Workshop.

HumanEgo helps robots learn tasks from first-person videos of human demonstrations, overcoming differences in how human hands and robot grippers look and move. With just 30 minutes of video per task, it achieved an average success rate of 92.5% across four tasks and transferred to different robots and environments without retraining.

Previously featured by UMIACS, the approach could make training data easier to collect by letting people demonstrate tasks themselves instead of remotely controlling a robot.

Shahidzadeh led “What to Attend, What to Keep: Skill-Conditioned Visuotactile Representation with Progress-Guided Event Memory,” which received the Best Paper Award at the Touch-to-Action: Enhancing Robot Manipulation through Tactile Perception Workshop.

The research examines how robots combine sight and touch to recognize a task’s progress. A robot screwing a light bulb into a socket, for example, must know when to stop. Successive turns can look almost identical, but increasing resistance signals that the bulb is fully tightened.

Different stages call for different sensory information: vision can guide reaching toward an object, while touch can reveal whether a grasp is secure or a movement is complete.

The team developed SkillFormer, a model that uses a language description of a skill to identify relevant sensory information and retains key observations from earlier steps. They tested it using demonstrations of lamp replacement, bottle capping and a search for a cube among boxes containing different shapes.

The cube search shows why memory matters. Touch helps identify a hidden object’s shape, while memory preserves which boxes have been searched and where the cube was found.

SkillFormer estimated task progress more accurately and detected key contact events sooner than comparison methods. During bottle capping, it reduced slip-detection delay by 87% compared with the progress-estimation models used as benchmarks.

A slip may last only a moment, then disappear from a system’s recent observations. SkillFormer retains that evidence as the task continues—a step toward helping robots recognize when to keep moving, when to adjust and when the job is done.

—Story by UMIACS communications group

Research Areas:

CS Undergraduate to Present AI Research at NeurIPS 2026

Samuel Malede Zewdu — Tue, 13 Oct 2026 13:10:10 +0000

Eric Gilerson’s single-author paper explores a method for performing exact arithmetic within language models.

October 13, 2026

When asked to multiply two numbers, a language model may predict the answer based on patterns in its training data rather than perform the calculation directly. University of Maryland undergraduate Eric Gilerson is investigating a way to change that.

Gilerson, a junior computer science major, has had a single-author paper accepted for a poster presentation at the 2026 Conference on Neural Information Processing Systems, known as NeurIPS. He will present his findings in Sydney, Australia, in December.

The conference brings together researchers in artificial intelligence and machine learning. For Gilerson, the acceptance marks his first experience submitting research to an academic conference.

“Having my first-ever research submission accepted to NeurIPS 2026 as an undergraduate is an incredible milestone that feels surreal and deeply rewarding,” Gilerson said.

His paper, “Don’t Learn What You Can Compute: Arithmetic Residual Blocks for Exact Arithmetic in Transformers,” examines how language models can perform calculations directly rather than rely on predictions that may produce incorrect results.

The approach adds an arithmetic component to the model’s architecture, letting it calculate answers without relying entirely on learned patterns.

“To solve this, I engineered an arithmetic residual block that calculates exact arithmetic results directly inside the transformer, injecting the answer into the token stream rather than forcing the model to approximate it,” Gilerson said.

In one experiment, the modified model achieved 99.95% accuracy across 36,000 short arithmetic expressions, compared with 4.56% for the unmodified model under the same testing conditions. Further research is needed to determine how the approach performs on more complex calculations and larger models.

Beyond raw calculation accuracy, Gilerson uncovered a broader benefit: parameter reallocation. By relieving the neural network from memorizing or approximating arithmetic, the model preserved parameter capacity, allowing it to learn unrelated tasks significantly faster and achieve higher benchmark performance.

“When you stop forcing a model to use representational capacity on deterministic rules such as arithmetic that computation can execute perfectly, you can free up its parameters for more complex, high-level reasoning,” Gilerson said. “My experiments showed that offloading arithmetic didn’t just eliminate computational errors; it freed the network to learn unrelated tasks faster and more effectively.”

Developing the system also required Gilerson to demonstrate its performance through experiments and respond to questions raised during the conference’s peer-review process.

“The review and rebuttal cycle was definitely intense, but I really appreciated how the reviewers challenged my initial assertions because it pushed me to design deeper experiments and empirically validate claims that had previously only seemed sound conceptually,” he said.

Gilerson describes himself primarily as a software engineer interested in building practical systems. That perspective also shapes how he thinks language models should handle tasks requiring precise calculations.

“Much like how humans naturally reach for a calculator or spreadsheet instead of wasting brain power on manual long division, language models should delegate mechanical steps,” Gilerson said.

Although the current implementation focuses on arithmetic, Gilerson sees potential applications in other rule-based operations, including unit conversions, calendar calculations and symbolic transformations.

His next steps include testing the architecture under different conditions and investigating whether language models can learn to formulate mathematical expressions that the component can solve directly.

As he prepares for NeurIPS, Gilerson hopes to use the poster presentation to discuss his findings and receive feedback from researchers working on related problems.

“I am now incredibly excited to head to Sydney to present the poster and discuss these findings directly with other researchers in the area,” Gilerson said.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

High Performance and Scientific Computing

Yafan Huang Explores Efficient, Reliable Computing Systems

Samuel Malede Zewdu — Thu, 08 Oct 2026 14:11:23 +0000

Huang develops methods to reduce data demands and address hardware errors in scientific computing and artificial intelligence systems.

October 08, 2026

As scientific facilities and artificial intelligence systems process growing amounts of data, researchers must contend not only with computing speed but also with the cost of moving and storing information and the possibility of hardware errors affecting results.

Yafan Huang, who joined the University of Maryland’s Department of Computer Science as an assistant professor in fall 2026, studies those challenges through high-performance computing, or HPC.

“I am very excited to join the Department of Computer Science,” Huang said. “UMD has a strong tradition in high-performance computing, including longstanding connections with the U.S. Department of Energy’s national laboratories, which is particularly appealing given my research background.”

Huang said the department’s work across computer systems and artificial intelligence also creates opportunities to connect areas that increasingly rely on the same computing infrastructure.

“I am also excited about the department’s strengths in both systems and AI, as I believe bringing these communities together will create valuable opportunities for research on efficient and reliable AI infrastructure,” he said.

That intersection is central to Huang’s research. He describes his broader research direction as DEFT, short for Data-Efficient and Fault-Tolerant computing.

“My research focuses on HPC, with the goal of making computing systems more Data-Efficient and Fault-Tolerant, an overarching direction I call ‘DEFT,’” Huang said.

The data-efficiency side of his work examines ways to reduce the resources required to move, store and access information. Huang develops high-performance data compression techniques and methods for computing systems that combine different types of hardware.

His current projects include data-reduction methods for such systems and approaches to managing information produced by large scientific facilities. One focus is the Advanced Photon Source at Argonne National Laboratory, where researchers use powerful X-rays to study materials and biological systems.

Huang has worked closely with Argonne researchers since 2021. His collaborations have also included scientists in light source science and seismic imaging, connecting his systems research to fields that generate large amounts of experimental and computational data.

“By reducing these overheads, my research aims to help scientists process more data within existing computing resources, enabling faster analysis and more efficient use of large-scale computing facilities,” Huang said.

The fault-tolerance side of DEFT addresses a different systems problem. Hardware can occasionally introduce errors during computation, potentially affecting a result even when an application continues to run. Huang studies software-based methods to detect and mitigate those errors without changing the underlying hardware.

His current work examines how such errors propagate through scientific applications and large language model workloads. He said improving reliability becomes increasingly important as researchers depend on complex computational systems to interpret scientific data and run AI models.

Huang earned his Ph.D. in computer science from the University of Iowa in 2026. He received the 2025 ACM-IEEE CS George Michael Memorial HPC Fellowship and was named a 2026 MLCommons ML and Systems Rising Star.

At UMD, Huang co-directs the Parallel Software and Systems Group (PSSG) with Professor Abhinav Bhatele. He plans to build on DEFT by bringing together HPC, computer systems and AI. His future research includes reducing memory use and data movement during AI inference, exploring compressed data representations during computation and studying how hardware errors affect modern AI workloads.

“I look forward to collaborating with colleagues and students across systems, AI, and computational science to develop computing techniques that can support both emerging AI applications and large-scale scientific discovery,” Huang said.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Alan Liu Helps Lead $17.3M NSF Programmable Cloud Lab Node for Biomanufacturing

Samuel Malede Zewdu — Tue, 06 Oct 2026 13:24:15 +0000

The four-year initiative will use AI and automation to support remotely accessible biomanufacturing experiments, with Liu co-leading the lab's data and AI systems.

October 06, 2026

Biomanufacturing underpins products ranging from vaccines to cancer therapies, but developing those products can require experiments that are time-consuming, costly and difficult to reproduce. A new University of Maryland-led project is bringing together artificial intelligence, robotics and real-time biological measurements to examine how those experiments can be designed and carried out differently.

UMD Department of Computer Science Assistant Professor Alan Zaoxing Liu is a co-principal investigator on the Collaborative for the Realization of Autonomous Biomanufacturing (CRAB) Lab, a four-year project supported by a $17.3 million award from the National Science Foundation. The lab will be housed at the Institute for Bioscience and Biotechnology Research (IBBR) in Rockville, Maryland, and will be remotely accessible to researchers across the country.

The project is led by principal investigator William Bentley, the Robert E. Fischell Distinguished Professor and director of the Robert E. Fischell Institute for Biomedical Devices. Liu serves as a co-principal investigator alongside Gregory Payne, Martha Wang, and Hussain Dahodwala. Ang Li from Electrical and Computer Engineering co-leads the AI efforts.

“This is a tremendously exciting moment,” Liu said. “I've been building the systems that make AI fast and reliable, and this award is a chance to apply that work to something as tangible as making medicines.”

The CRAB Lab is one of 20 Programmable Cloud Laboratory Node sites supported through an NSF initiative to develop AI-enabled laboratories for automated science and engineering. Researchers will be able to program workflows that design, run and analyze biomanufacturing experiments while automated instruments and AI systems collect data and help determine what steps to take next.

Liu’s group is responsible for the project’s data and AI backbone. That work includes managing the large volumes of measurement data generated by laboratory instruments, coordinating experiments from beginning to end and training models to recognize what Liu describes as “biological vital signs,” or patterns that can indicate whether a biomanufacturing process is operating as expected or beginning to drift.

For Liu, the project extends research he has pursued in computer systems, networking, and large-scale data infrastructure. His research at UMD includes designing scalable and reliable systems for processing and monitoring data. The CRAB Lab will apply that work in an environment where computing systems interact directly with physical experiments.

“Building AI that actually runs a lab, not just a benchmark, is exactly the kind of challenge I hoped to take on,” Liu said. “I'm grateful to NSF, and to William Bentley and the IBBR team for bringing computer science into the effort from day one.”

The project will use bioelectronic measurement tools developed through UMD partnerships with the National Institute of Standards and Technology (NIST) and the Food and Drug Administration (FDA). The tools are designed to produce high-volume, high-speed measurements for training AI models. Those models could identify combinations of measurements associated with product quality and other process characteristics, giving the automated system information it can use as an experiment proceeds.

“Almost everything in modern medicine, from vaccines to cancer therapies, depends on biomanufacturing, but the experiments behind it are slow, expensive and often hard to reproduce,” Liu said. “We're building AI agents that can design an experiment, hand it to the robots, monitor the data in real time and decide what to try next, so a scientist can log in from anywhere and have results the next morning.”

Remote access is also intended to expand access to automated biomanufacturing equipment. The lab is expected to initially serve 30 biopharmaceutical companies through the Advanced Mammalian Biomanufacturing Innovation Center. Bentley said what excites him most is that the lab is “contributing to a national resource” that researchers across academia, industry and government can draw on.

“That shortens the path from idea to result by months,” Liu said. “Just as important, a graduate student at a small school will have the same AI-driven lab at their fingertips as a large company, and that kind of access changes who gets to do the science.”

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

UMD Computer Scientist Heng Huang Receives Funding to Develop Wearable AI Vitals-Monitoring Device and AI-Enabled Medical Imaging Test Bed

Samuel Malede Zewdu — Mon, 05 Oct 2026 15:21:19 +0000

The projects aim to predict hospital readmissions in high-risk heart failure patients and establish reliable methods for evaluating the safety and performance of AI-enabled medical imaging systems.

October 05, 2026

University of Maryland computer scientist Heng Huang is part of two new federally funded projects aimed at improving the use of artificial intelligence in healthcare: a four-year, $1.2 million award to develop a wearable device that predicts hospital readmission in high-risk heart failure patients and a $750,000 contract to create a platform for evaluating AI-enabled medical imaging systems.

“AI has so much potential to help us address human health problems, and developing these kinds of real-world tools is one important way forward,” said Huang, the Brendan Iribe Endowed Professor of Computer Science at UMD.

With the $1.2 million grant, funded jointly by the National Institutes of Health and the U.S. National Science Foundation, Huang will collaborate with Wei Gao at the University of Pittsburgh to develop and test a small, affordable wearable device equipped with AI. The device will use magnetic sensing and other sensors to monitor heart failure patients’ vital signs at home. New machine learning models will combine the data in real time to predict patients’ risk of hospital readmission.

“Heart failure is the leading cause of hospitalization in older adults, and hospital readmissions after discharge are common and have become the top reason for worse clinical outcomes,” said Huang, who also has a joint appointment in the University of Maryland Institute for Advanced Computer Studies and leads applied AI at the University of Maryland Institute for Health Computing (UM-IHC). “The problem is especially serious among patients with obesity, who are 30% more likely to require rehospitalization than other heart failure patients.”

The models will enable the researchers to identify risk factors from patients’ everyday vital signs and important biomarkers and predict readmission before it occurs.

“Our ultimate goal is to prevent those hospital returns to help save lives,” Huang said.

Data collected by the device could also help the researchers better identify the mechanisms and symptoms of heart failure and support clinicians in making an initial diagnosis.

Assessing AI-enabled medical imaging

Huang is also collaborating with UM-IHC colleagues on a new Food and Drug Administration (FDA)-funded project to create a platform to evaluate AI-enabled medical imaging systems for detecting pulmonary embolism, a condition in which a blood clot blocks an artery.

“This project goes to the question of the trustworthiness of clinical AI in patient care. If we’re going to use AI to review radiology scans, we need to be able to evaluate what it gets right and what the failure points are,” said lead awardee Florence Doo, an assistant professor in diagnostic radiology and nuclear medicine at the University of Maryland School of Medicine (UMSOM) who co-leads with Huang the AI-enabled medical imaging team in UM-IHC’s Center for Applied AI. “We are building a test bed, or dashboard, that will let clinicians and the FDA evaluate post-market performance and safety of AI in very precise ways, in a health system and at scale.”

The researchers will first standardize hospital imaging data, radiology reports and hospital reports for use in evaluating medical AI. Those reports will come from nearly 6,000 patient records from the University of Maryland Medical System as well as public datasets.

“We will then deliberately stress-test AI models to see where they fail and document what kinds of mistakes AI makes and under what circumstances,” Huang said.

The team also includes UMD Computer Science Professor Adam Porter and Melvin Sharoky MD Professor of Medicine at UMSOM Bradley Maron—the co-executive directors of the UM-IHC.

“It’s not enough to just validate an AI-enabled device on benchmark data at a single point in time; we must also validate that the device works correctly throughout its useful lifetime,” Porter added. “Therefore, post-market evaluation of AI devices is a key focus of this project.”

—Story by CMNS

###

This work is supported by the National Institutes of Health and the U.S. National Science Foundation (Award No. R01NR022545) and the FDA (Contract No. 75F40126C00094). This article does not necessarily reflect the views of these organizations.

Research Areas:

Teaching Robots to Follow the Motion

Samuel Malede Zewdu — Mon, 05 Oct 2026 14:50:14 +0000

October 05, 2026

Illustration by iStock

Watch someone pour almonds into a bowl, and the action is easy to follow: a container moves, tilts and releases its contents. Teaching a robot to do the same can require costly demonstrations specifying how its joints and gripper should move.

University of Maryland researchers have developed an artificial intelligence model that learns from videos how hands, tools and objects move and then uses those patterns to help robots perform tasks.

Called μ₀, or Mu Zero, it is a type of AI called a world model, which predicts how a scene will change during an activity. Given an image and an instruction, such as “place the cup in the sink,” Mu Zero predicts how selected points on objects will move through three-dimensional space.

In tests with a robot arm, the system outperformed several competing approaches at placing objects in a sink, pouring almonds and unfolding a towel.

The team includes UMD computer science associate professors Furong Huang and Jia-Bin Huang and collaborators at UMD and Seoul National University. Both professors hold appointments in the University of Maryland Institute for Advanced Computer Studies (UMIACS). Furong Huang is also a core faculty member in the UMD Center for Machine Learning.

Their paper, “μ₀: A Scalable 3D Interaction-Trace World Model,” earned the second-place Best Paper Award at the IROS 2026 RoBoWoMo workshop in late September. It has also been accepted for presentation in November at the Conference on Robot Learning.

The project addresses a stubborn obstacle: videos of physical activities are plentiful, but recordings paired with precise robot commands are expensive to collect. Those commands are also hardware-specific; instructions for one robot’s arm may not translate to another.

Some AI approaches predict what happens next by generating future video frames, devoting substantial computing capacity to recreating appearances and backgrounds. Others predict robot commands directly, tying their training to specific machines.

Mu Zero focuses on movement, representing the paths of selected points on hands, tools and objects—including where they make contact—as smooth curves called traces.

“Think of it as tracing the choreography of a task. When a cup moves toward a sink, the useful information includes where the cup travels and how it turns,” said Seungjae “Jay” Lee, a UMD computer science doctoral student and co-lead author of the paper. “The model represents that movement with smooth curves, preserving spatial information without reconstructing every pixel.”

Those visible paths also help researchers inspect the model’s predictions, added Lee, who is co-advised by Furong Huang and Jia-Bin Huang.

To create training data, the team developed TraceExtract, a system that converts human and robot videos into motion traces paired with descriptions. It selects meaningful points, tracks their movement in three dimensions and pairs individual motion events with language descriptions. It also distinguishes object movement from camera movement.

Mu Zero learns to forecast these traces. A separate component, called an action expert, learns to translate information from the model into commands for a specific robot.

Mu Zero’s initial training on video requires no accompanying robot commands, although the action expert still learns from robot demonstrations. The world model remains fixed during that adaptation, allowing its motion knowledge to be reused.

“Our approach separates learning what needs to move from learning how a particular machine can make that movement happen,” said co-lead author Yoonkyo Jung, a UMD doctoral student in electrical and computer engineering advised by Furong Huang.

In physical experiments, a robot arm with a two-finger gripper achieved an average success rate of 91.7% across three tasks, exceeding two competing systems by 20 and 11.7 percentage points, respectively.

Simulation results were mixed. Across eight kitchen tasks, Mu Zero achieved a 30.25% average success rate, ahead of one competitor’s 25.25% but below the other’s 42%. It also outperformed earlier methods for predicting motion traces in two and three dimensions.

“This work contributes to physical intelligence—AI systems that connect perception and language with movement in the world,” said Furong Huang. “Learning reusable motion patterns from videos could reduce dependence on costly, hardware-specific training data.”

She noted that the system still has limits. Tracking and captioning errors can affect learning, and the model does not explicitly account for forces or touch. Its performance on a broader range of robots and tasks remains untested.

The experiments nevertheless demonstrate a useful division of labor, she explained: videos teach the model how objects move, while robot demonstrations teach a specific machine how to carry out those movements. That separation allows the system to learn from recordings that contain no detailed robot commands.

—Story by UMIACS communications group

Research Areas:

Natural Language Processing

UMD CS Students Connect With Employers at Career Fair

Samuel Malede Zewdu — Fri, 02 Oct 2026 13:31:57 +0000

The Fall 2026 event brought 1,169 students together with representatives from 59 employers seeking candidates for internships and full-time positions.

October 02, 2026

The University of Maryland Department of Computer Science brought 1,169 students and representatives from 59 employers together at its Fall 2026 Career and Internship Fair on Sept. 28. Held at the Adele H. Stamp Student Union, the event gave students direct access to recruiters seeking candidates for internships and full-time positions in computer science and related fields.

For students, the fair offered a chance to speak directly with employers and better understand how their coursework and experiences translate to the workplace.

Senior Sofia Santiago, who is pursuing a double degree in computer science and social data science, said the event gave her time to have more detailed conversations with recruiters.

“My experience at the career fair has been very positive,” Santiago said. “I’ve been able to talk to all the companies and recruiters that I intended to talk to, which is a win.”

Santiago shared that one conversation with representatives from Appian highlighted the importance of technical preparation and involvement outside the classroom.

“They’re looking for students who are involved in student organizations on campus, and they’re also looking for students who have the technical skills in the targeted positions,” Santiago said. “If you’re doing software engineering, they want to know your coding background, projects you’ve done and also how you’re implementing your software engineering skills outside of school.”

She also saw the fair as an opportunity to expand her professional network and learn about employers she hadn't previously considered.

“The fair has helped me learn what opportunities are out there in the first place and companies that are not necessarily known or super big,” Santiago said. “I see this year a lot of new companies that I didn’t see last year.”

Junior computer science major Bishop Daheu attended the fair to explore and connect with employers.

“I’m just looking for career opportunities, looking for growth and a way and a means to apply my skills that I’ve gathered up to this point,” Daheu said.

Daheu said he is particularly interested in government technology and government contracting, including software engineering, cloud engineering and solutions architect roles. He prepared by researching companies’ missions, values, projects and technologies before speaking with recruiters.

That preparation reflects what some employers said they look for when meeting students.

Maggie Neff, a senior data scientist at Fannie Mae and a UMD alumna who studied mathematics and economics, attended the fair to recruit students for data science and software engineering opportunities. Fannie Mae representatives were recruiting for full-time and internship positions in its Single Family Analytics group, along with technology-focused software engineering internships.

Neff said candidates may come from a range of academic backgrounds, but employers look for qualities that apply across disciplines.

“The underlying factor for all of those is intellectual curiosity and an analytical mindset that can be applied in a lot of ways in a lot of different roles,” Neff said.

She said students can demonstrate those qualities through internships, research and class projects. Resumes should explain the tools and programming languages they used, the problems they worked on and, when possible, the measurable results.

Neff added that researching an employer and asking informed follow-up questions can also help students during career fair conversations and interviews.

“It’s nice to see that someone has done some due diligence to look into what we do broadly as a company, but then also in the specific role,” Neff said. “I think it’s cool when people want to hear about our favorite parts of the job or projects that we’ve worked on and have follow-up questions that show engagement.”

A recent survey by the University Career Center shows that UMD’s Department of Computer Science achieved a 95% job placement rate for the Class of 2025, which includes graduates from August and December 2024 and May 2025. The rate includes graduates who were employed, continuing their education or pursuing other post-graduation opportunities.

—Story by Samuel Malede Zewdu, CS Communications

Building Disney+, ESPN+ and a New SiriusXM

Samuel Malede Zewdu — Thu, 01 Oct 2026 14:50:34 +0000

Jared Wolinsky (B.S. ’09, computer science) credits his University of Maryland education and the QUEST Honors Program for preparing him for his current role as group vice president of software engineering at SiriusXM.

October 01, 2026

Jared Wolinsky (B.S. ’09, computer science) leads SiriusXM's engineering efforts to build next-generation apps. Photo courtesy of Jared Wolinsky

He started tinkering with computers in elementary school. Using an old IBM machine his grandfather passed down to his family, Wolinsky learned Visual Basic—an early programming language developed by Microsoft for Windows.

At first, he built apps to play pranks on his friends through AOL’s messaging platform.

“I had a little mischievous streak at the time,” he said, laughing. “It was never anything nefarious—just doing the things that would sound fun for a middle schooler.”

But his love for coding grew into something bigger. He sought resources to develop his skills and found like-minded people in school. By the time he got to the University of Maryland, he chose to major in computer science and then joined the QUEST Honors Program, where he sharpened his skills at the intersection of technology and business.

“I thought maybe I would start a company or work on the business side in some way, so I just wanted exposure to that,” Wolinsky said. “The collaboration between the business school and computer science really appealed to me.”

Now, Wolinsky is living out his childhood dream as the group vice president of software engineering for the experience engineering group at SiriusXM, whose automotive app is available on over 80% of new cars sold in the United States. He leads engineering for the listener experience across SiriusXM and Pandora, spanning mobile, web, connected devices and the next-generation in-vehicle app. His team develops the app’s functionality for playback, browsing, search, personalization and everything else related to a user’s listening experience.

“I’m really happy with where I am at the moment,” Wolinsky said. “I just want to continue working on really high-leverage, interesting consumer products.”

Before arriving at SiriusXM, Wolinsky worked at MLB Advanced Media, where he helped build a platform for fans to stream baseball games. That platform also hosted NHL games, Major League Soccer games, Eurosport, HBO Now, Sony PlayStation Vue and other direct-to-consumer video streaming products. As a content engineer, Wolinsky worked with a team on the back end to configure the content pipeline and content catalog. Their system allowed the platform to pull metadata and artwork to inform how users manage, curate, search and discover content.

The technology eventually became the foundation of BAMTech, a streaming technology company spun out of MLB Advanced Media. Disney acquired a minority stake in 2016 and later took majority ownership, rebranding the organization as Disney Streaming Services. At BAMTech and later Disney Streaming, Wolinsky and his teams helped build the technology behind ESPN+ and Disney+.

“That was definitely one of the highlights of my career,” Wolinsky said.

Then in 2022, Wolinsky joined SiriusXM for the opportunity to apply what he learned building large-scale streaming products to another well-known consumer entertainment brand. As a longtime SiriusXM and Pandora listener, the opportunity was especially appealing. He joined SiriusXM as vice president of platform engineering, helping build the technology foundation and developer platforms used by the company’s engineering teams. He then became vice president of software engineering, leading the backend services platform, and finally reached his current role as group vice president responsible for experience engineering.

Day to day, Wolinsky is “in one big meeting,” he jokes. He spends much of his time on organizational design, people leadership, technology strategy and long-term planning. Much to the chagrin of his middle school self, he does almost no computer programming. Still, he sees similarities between his developer roles and his current leadership role.

“I consider myself an organizational engineer,” he said. “I love figuring out the perfect way that our team should be structured to set us up for success. It is a little bit like software engineering—fitting the pieces in the right place.”

In part, he credits his experiences at UMD and in QUEST for preparing him for success at the intersection of entrepreneurship, technology and business. For QUEST’s capstone projects, students work on interdisciplinary teams solving projects for real-world companies.

“The entrepreneurial aspect and real-world application really did it for me. I really liked the idea of working with actual clients and seeing how the stuff that we were learning in class applied,” he said. “I also think that being part of a team serving a real client in a semi-real environment sets you up for industry better than going straight from the classroom.”

And as much as Wolinsky misses coding, he gets as much satisfaction leading teams as he did developing software himself.

“The thrill of creating something is pretty apparent to anybody who's done it. Seeing working software take shape after it was just an idea in your mind—there’s nothing like that. It's why anyone who builds things for a career gets into it,” Wolinsky said. “But there's something equally satisfying about seeing your team deliver something amazing that you had almost no part in actually building yourself. That you were responsible for setting the conditions that made it possible for that to happen. That brings a different type of satisfaction.”

—Story by CMNS

In Pictures: 2026 Tech Block Party

Samuel Malede Zewdu — Wed, 30 Sep 2026 16:48:09 +0000

Computing Catalyst’s annual event brought students together with campus organizations, UMD programs and industry representatives at Hornbake Plaza.

September 30, 2026

About 400 students gathered at Hornbake Plaza on Sept. 25 for Computing Catalyst’s annual Tech Block Party, where they connected with student organizations, UMD programs and technology companies.

The images below capture scenes from the event and the connections students made throughout the afternoon.

Previous

Next

—Portrait Article by Samuel Malede Zewdu, CS Communications

What Keeps People Going in Dark Times? New Research Finds Unexpected Answers

Samuel Malede Zewdu — Tue, 29 Sep 2026 20:25:12 +0000

Philip Resnik and Rebecca Resnik used computer-assisted analysis to examine more than 16,000 personal accounts describing reasons people choose to continue living.

September 29, 2026

What can thousands of personal stories tell us about the reasons people keep going during their darkest moments?

University of Maryland computational linguist Philip Resnik and clinical psychologist Rebecca Resnik recently joined CBS News correspondent Natalie Brand to discuss new research examining that question.

Philip Resnik, who is also an affiliate professor in the Department of Computer Science, is the lead author of the forthcoming Scientific Reports paper “16,648 Reasons to Live Instead of Dying by Suicide: Insights from a Computer-Assisted Content Analysis.” A professor of linguistics with a joint appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS), he leads research combining machine learning, computational linguistics and human expertise to study complex questions in mental health.

The research team used machine learning to sift through more than 16,000 firsthand accounts and identify recurring patterns, which experts then interpreted and organized. The analysis highlighted themes including concern for loved ones and pets, everyday pleasures such as music and television, hope and a sense of meaning and even emotions such as fear, anger and spite.

For Rebecca Resnik, a clinical psychologist who collaborated with her husband on the project, the findings also have potential relevance to clinical care. Paying attention to seemingly small sources of connection and anticipation can provide important insight into what matters to an individual.

The CBS interview offers a closer look at the researchers’ work and at how new computational approaches are helping expand the study of mental health.

Research Areas:

UMD CS Announces Hatchery Teams for 2026-2027

Samuel Malede Zewdu — Tue, 29 Sep 2026 13:46:47 +0000

The five student-led startups focus on job searches, home-service calls, neighborhood equipment sharing, engineering workflows and passive health monitoring.

September 29, 2026

The University of Maryland’s Department of Computer Science has selected five student teams, Pinloop, NexFlow, Shelf, Simulace and PulseFi, for the 2026-2027 academic year of the Mokhtarzada Hatchery Program.

The Hatchery provides selected teams with up to $10,000 in funding, access to startup mentors and workspace in the Brendan Iribe Center for Computer Science and Engineering. Teams use the academic year to develop and test ideas, identify potential users and refine products and business models.

Funded by UMD alumni and siblings Haroon (B.A. ’01, economics), Idris (B.S. ’10, computer science) and Zeki Mokhtarzada (B.S. ’01, computer science), the program was established in 2021 to give student founders a structured environment to explore and test ideas.

Each team is expected to present its progress at the program’s annual showcase at the end of the spring semester.

Pinloop

Pinloop is developing a platform to reduce the time spent searching for and applying for jobs. The system combines a job database with tools that allow artificial intelligence agents to search listings, complete application forms and track outcomes.

The platform is also designed to collect information about employer responses and the types of applications or outreach that lead to interviews or offers. The team says those results could help users identify employers and application approaches associated with higher response rates.

Pinloop is led by UMD undergraduate Andrew Allen (B.S. ’28, computer science).

NexFlow

NexFlow is developing an AI voice receptionist for small home-service businesses, including HVAC, plumbing and locksmith companies. The system is designed to answer calls when owners or technicians are unavailable, respond to basic questions about services and pricing and book appointments into a business’s calendar.

According to the team, the system can also identify urgent calls, such as reports of a burst pipe or a lack of heat, and escalate them to the owner based on customized rules.

The NexFlow team includes UMD undergraduate Aditya Kaushik (B.S. ’30, computer science) and Northeastern University student Deepanshu Kaushik (B.S. ’28, computer science).

Shelf

Shelf is a neighborhood marketplace for borrowing equipment that owners may use only occasionally, including ladders, leaf blowers and pressure washers. Users can list items and make them available to nearby residents for short-term use.

The platform uses Stripe for payments and identity verification. According to the team, Shelf is available through the App Store in 10 neighborhoods and includes listings across more than 15 item categories. The team aims to offer an alternative to buying equipment you may use infrequently.

The Shelf team includes UMD students Ranita Rajkumar (B.S. ’27, information science) and Arush Jain (B.S. ’27, computer science).

Simulace

Simulace is developing a software-neutral system intended to manage handoffs among engineering tools used in hardware design. Hardware teams may move models among computer-aided design, thermal, structural and optical software as they evaluate design changes.

The platform is designed to automate those transfers while working with software a company already uses and operating within the company’s network. Its developers say the system is intended to make each step reproducible while reducing manual data movement between programs.

The Simulace team includes UMD students Rowand Zara (B.S. ’29, aerospace engineering and applied mathematics) and Shreyan Pasham (B.S. ’29, computer engineering).

PulseFi

PulseFi is developing a passive Wi-Fi-based system intended to monitor health-related signals and movement patterns without requiring wearable devices or cameras. The system uses two wall-powered nodes and analyzes changes in Wi-Fi signals as they interact with a person’s body.

According to the team, the technology is being developed to measure signals including heart rate and breathing rate and to identify events or changes related to apnea, falls and gait. The team is focusing on potential applications involving older adults and neurodegenerative decline.

The PulseFi team includes UMD undergraduate Pranay Kocheta (B.S. ’30, computer engineering) and University of California, Santa Cruz doctoral student Nayan Bhatia (Ph.D. ’27, computer science and engineering).

—Story by Samuel Malede Zewdu, CS Communications

Computing Catalyst Forms Inaugural Advisory Board

Samuel Malede Zewdu — Fri, 25 Sep 2026 18:26:34 +0000

The eight-member body will provide industry perspectives, support student programming and strengthen connections among students, alumni and professionals.

September 25, 2026

Members of the University of Maryland Computing Catalyst’s inaugural advisory board are, top row from left, Anjelica Dortch, Kerron Duncan, Ume Habiba and Ruchi Kansal Gupta; and bottom row from left, Amisha Kunvar, Angel Plane, Awalin Sopan and Anna Yankova.

The University of Maryland Computing Catalyst has formed its inaugural advisory board, bringing together eight professionals from technology, government, consulting and engineering. The board met for the first time in late August.

Computing Catalyst provides mentoring, professional development, scholarships and community-building programs for UMD students in computing, along with outreach for local K-12 students. During the 2025-26 academic year, more than 300 UMD students and 116 local middle and high school students participated in its programs.

“Pursuing excellence in computing means creating opportunities for students to learn, grow and succeed,” said Department Chair Matthias Zwicker, who holds the Elizabeth Iribe Chair for Innovation and the Phillip H. and Catherine C. Horvitz Professorship. “I am deeply grateful for the advisory board’s engagement to help the Computing Catalyst advance that mission.”

Board members will serve two-year terms and act as an extension of Computing Catalyst staff by providing industry perspectives and feedback on programs, volunteering with students, drawing on professional networks and supporting fundraising efforts. Seven of the eight members are UMD alumni, and several previously participated in or volunteered with Catalyst programs.

“We are excited about this group because the board members have been eager to jump in and volunteer, help plan programs. They also have great ideas and energy to help fundraise for the Dr. Jan Plane Endowed Program Support Fund,” said Caitlin Rudy, program manager of Computing Catalyst.

The board brings experience across technology, government, consulting, engineering, cybersecurity and artificial intelligence.

Anjelica Dortch

Dortch (B.A. '11, philosophy) is founder and CEO of Vivid Strategies LLC, a government relations and strategic advisory firm focused on technology, cybersecurity, artificial intelligence and national security policy. Her career includes private-sector work and a decade of federal service, including roles in the Executive Office of the President.

She also earned a master’s degree in financial management and information systems from University of Maryland Global Campus.

Kerron Duncan

Duncan is an AI catalyst for defense and systems at GE Aerospace, where he leads artificial intelligence strategy, adoption and governance. He previously spent 25 years at Northrop Grumman in engineering, digital transformation, automation and technology leadership roles. He is also co-founder of Career Factory 360, a career development practice for STEM professionals and students.

Duncan earned bachelor’s and master’s degrees in electrical engineering from Morgan State University.

Ume Habiba

Habiba (B.S. ’24, information science) is a software engineer and employee creator at Microsoft. Her work and outreach interests include technology advocacy, education and career development.

While at Maryland, she attended the Grace Hopper Celebration with scholarship support connected to Computing Catalyst and served as a teaching assistant for three computer science courses.

Ruchi Kansal Gupta

Gupta (B.S. '94, computer science) is executive program director and partnerships lead for the Google Developer Community across New York City and founder of Monamedia, an advisory practice focused on brand strategy, artificial intelligence-driven growth and marketing.

Her career has included roles at IBM, Apple, HSBC, Sony Pictures Entertainment and Sesame Workshop, along with work involving virtual reality and artificial intelligence-based medical training.

During her time at UMD, she studied under faculty members including Jandelyn Plane, whose computing programs her daughters later attended.

Amisha Kunvar

Kunvar (B.S. ’22, computer engineering) is a senior technology consultant at PwC, where she supports large-scale technology transformations involving mergers and acquisitions integration, product operations, IT governance and investment planning.

She has also worked on education and career development initiatives, including mentoring UMD students through recruiting and professional development programs.

Angel Plane

Plane (B.S. ’15; M.S. ’17, computer science) is an AI and machine learning engineering leader and government contractor whose work includes cloud architecture, machine learning infrastructure and autonomous AI systems.

Her career has included software development and leadership of multidisciplinary engineering teams. She also volunteers with initiatives that expand access to technical fields.

Awalin Sopan

Sopan (M.S. ’11, computer science) is a Principal AI Engineer in Cyber Defense at Workday and an alumna of the University of Maryland, College Park, where she earned her Master's in Computer Science. With a career spanning Sophos, Mandiant, and public sector leadership as a White House Presidential Innovation Fellow, she builds intelligent platforms that bridge complex data and human workflows. Her connection to UMD is rooted in the Human-Computer Interaction Lab that shaped her passion for human-centered AI and design, and in serving the American public through a global perspective on responsible AI and Data. Joining the Catalyst board represents an exciting opportunity to give back to the institution that launched her journey, partnering with a community dedicated to innovation and impactful collaboration.

Anna Yankova

Yankova (MBA ’22, executive MBA) is a project manager at the Maryland Health Benefit Exchange, where she manages technology initiatives involving artificial intelligence, automation, data analytics and digital transformation.

She previously held leadership roles in the Baltimore Mayor’s Office and the Maryland Office of the Secretary of State. Her work included programs supporting immigrant and refugee communities, international partnerships and Maryland’s Sister States Program.

She also holds a master’s degree from American University’s School of International Service and a bachelor’s degree from Odesa State University in Ukraine.

—Story by Samuel Malede Zewdu, CS Communications

William Gasarch and Mohammad Hajiaghayi Co-Author Book on Algorithmic Limits

Samuel Malede Zewdu — Wed, 23 Sep 2026 17:05:54 +0000

Their new book, co-authored with MIT professor Erik Demaine, examines how computer scientists determine the limits of solving computational problems.

September 23, 2026

University of Maryland computer science professors William Gasarch, left, and Mohammad Hajiaghayi.

University of Maryland Department of Computer Science faculty members William Gasarch, a professor of computer science, and Mohammad Hajiaghayi, the Jack and Rita G. Minker Professor of Computer Science, have co-authored a forthcoming book examining how computer scientists determine how difficult problems are to solve and whether faster solutions are possible. Computational Intractability: A Guide to Algorithmic Lower Bounds, co-authored with Massachusetts Institute of Technology Professor Erik Demaine, is scheduled for publication on Oct. 13 by MIT Press.

The book focuses on algorithmic lower bounds, which researchers use to understand the limits of how efficiently a computational problem can be solved. Rather than focusing only on finding faster algorithms, the book examines how researchers can determine when certain limits cannot be overcome.

The project grew out of conversations between Hajiaghayi and Demaine in 2013 about how computational complexity was taught. They saw an opportunity to approach the subject from the perspective of researchers who design and analyze algorithms.

“We felt there was not a class that taught complexity to students from the perspective of people working and thinking on algorithms,” Hajiaghayi said.

That discussion led them to develop related graduate courses at UMD and MIT in fall 2014. The courses examined the limits of solving a variety of computational problems and eventually produced lecture notes and other materials that became the book's foundation.

Hajiaghayi and Demaine later invited Gasarch, whose research includes computational complexity theory, to join the project.

Hajiaghayi said the courses examined lower bounds across different types of algorithms, providing the framework that later became the book's basis.

“We felt a book in this area would be unique, and we made it happen,” Hajiaghayi said.

For Gasarch, the book centers on a fundamental question in computational complexity: How difficult is a given problem to solve?

“Every chapter classifies a set of problems in terms of how hard they are,” he said.

The book covers several areas of computational complexity, including the limits of parallel computing and NP-completeness, while emphasizing specific problems and the methods researchers use to analyze them.

Through those examples, the authors aim to give readers a broader understanding of not only how algorithms solve problems, but also how computer scientists determine the limits of what those algorithms can do.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Algorithms and Theory

UMD Computer Science Undergraduate Program Ranks in the Top 20 for the 7th Consecutive Year

Samuel Malede Zewdu — Tue, 22 Sep 2026 13:58:12 +0000

The program ranks in the Top 10 among public institutions overall and for the specialties of artificial intelligence and cybersecurity, according to U.S. News & World Report.

September 22, 2026

The University of Maryland’s undergraduate computer science program ranks No. 17 overall in the 2027 edition of U.S. News & World Report’s “Best Colleges.” UMD ranks in the Top 10 among public institutions overall and for the specialties of artificial intelligence and cybersecurity.

“Undergraduate computer science programs have only been ranked by U.S. News & World Report for the past seven years, and the University of Maryland has ranked in the top 20 every year," said Matthias Zwicker, chair of UMD’s Department of Computer Science and holder of the Elizabeth Iribe Chair for Innovation and the Phillip H. and Catherine C. Horvitz Professorship. “Above all, this is a testament to the outstanding commitment to excellence of our students and faculty. I couldn't be more proud of their achievements.”

UMD boasts one of the largest computer science programs in the country, with 3,000 undergraduate computer science majors who can pursue a specialization in cybersecurity, data science, machine learning or quantum information. UMD ranks No. 1 among non-online U.S. universities for conferring the most undergraduate degrees in computer and information science, according to The Chronicle of Higher Education.

The 2025 Graduation Survey administered by the University Career Center reported a 95% placement rate of UMD’s computer science graduates with a median starting salary of $119,000 and a range of notable employers, including Amazon, Capital One, Meta, Bloomberg, Google and Microsoft. Of the survey respondents, 78% reported full-time employment and 16% continued their education.

“The University of Maryland's undergraduate computer science program gives students the depth and breadth to lead in the technologies shaping our future, from artificial intelligence to cybersecurity,” said Amitabh Varshney, dean of UMD’s College of Computer, Mathematical, and Natural Sciences and a professor of computer science. “We are proud to be consistently recognized among the nation’s best for preparing the next generation of innovators and leaders in computing.”

Earlier this year, the Department of Computer Science announced plans to launch a new Bachelor of Science degree in AI in Fall 2027. In this program, students will learn to build AI systems and algorithms from the ground up, understanding the training of models from data and systems perspectives, and then use those techniques in interdisciplinary applications.

Graduates of UMD’s undergraduate computer science program include: Michael Antonov ’03, Oculus co-founder; David Baggett ’92, Inky founder; Sergey Brin ’93, Google co-founder; Katherine Calvin ’03, former NASA chief scientist; Anthony Casalena ’05, Squarespace founder; Jonathan Chen ’14, Nitra founder; Judith Dotson ’85, retired Booz Allen Hamilton president of the global defense sector; Josh Kamdjou ’14, Sublime Security founder; Akash Magoon ’18, Adonis co-founder; Ruvi Kitov ’97, Tufin co-founder; Cassandra Lewis ’09, Rogue Studios London founder; Andy Looney ’86 and Kristin Looney ’88, Looney Labs co-founders; Idris Mokhtarzada ’10 and Zeki Mokhtarzada ’01, Truebill co-founders; Sujal Patel ’96, Nautilus Biotechnology co-founder; Andrew Reisse '01, Oculus co-founder; Jagdeep Singh ’86, Rhoda AI founder; and Adam Wenchel ’99, Arthur co-founder.

UMD’s graduate computer science program ranks 12th overall (No. 7 among public institutions) according to U.S. News & World Report.

—Story by CMNS

One Robot, Many Bodies: New System Helps Four-Legged Robots Adapt

Samuel Malede Zewdu — Mon, 21 Sep 2026 16:14:05 +0000

Dinesh Manocha and collaborators developed the MorFiC framework, which allows a locomotion policy trained on one robot design to adapt to four-legged robots with different weights, shapes and motors.

September 21, 2026

A four-legged robot can learn to run and recover from a shove through reinforcement learning, a form of artificial intelligence in which a system learns through trial and error by receiving rewards for useful behavior. But transfer that robot’s control system into a body with a different weight, shape or set of motors, and those skills can fall apart.

University of Maryland and George Mason University researchers have developed a more efficient approach. Their framework, the Morphology-aware FiLM Critic (MorFiC), trains a locomotion policy on one robot design and applies it to others without retraining. A paper describing the work has been accepted to the 2026 Conference on Robot Learning (CoRL 2026) in Austin, Texas.

“Our approach trains on a single robot in about two hours,” said the study’s lead author, Prakhar Mishra, who earned a Master of Engineering in Systems Engineering at UMD with a specialization in robotics. Comparable approaches can require tens or hundreds of hours because they train on many robot designs or use much larger models, he said.

Mishra led the project’s conceptual development, methodology, implementation, experimental design and analysis. George Mason computer science doctoral student Amir Hossain Raj contributed to the experimental implementation, robot deployment and evaluation. UMD’s Dinesh Manocha and George Mason’s Xuesu Xiao supervised the work and provided technical guidance and manuscript review.

Manocha is a Distinguished University Professor of Computer Science with an appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS) and served as Mishra’s adviser. Xiao, an assistant professor of computer science, is Raj’s adviser.

Much of the UMD research took place in Manocha’s Geometric Algorithms for Modeling, Motion and Animation (GAMMA) Lab.

The researchers trained MorFiC on a single quadruped—a four-legged robot—rather than a collection of robot bodies, then tested the same policy on seven platforms. This “zero-shot” transfer required each machine to use what the system had already learned, without further training or adjustments.

Differences in mass, leg geometry, joint limits and motor strength change how a robot responds to commands. A policy—the learned rules translating sensor readings into actions—may work smoothly on its original robot but become unstable on another.

MorFiC focuses on the part of reinforcement learning that evaluates those actions. In an actor-critic system, the “actor” decides what the robot should do, while the “critic” estimates how useful that action will be and guides the actor’s learning.

The researchers found that the critic can become poorly calibrated when working across different robot bodies. The same pose or action may produce different results depending on the machine, yet a conventional critic may average those outcomes and steer learning in the wrong direction. MorFiC instead supplies the critic with information such as the robot’s mass and joint limits, allowing it to judge an action in the context of the body performing it.

In simulations, MorFiC produced the highest forward speed among the tested approaches on five of seven platforms. On an AlienGo robot, it reached 1.98 meters per second, while conventional reinforcement-learning variants remained below 0.65 meters per second. AlienGo differed substantially from the Unitree Go2 design used for training.

The researchers evaluated speed alongside “survival,” or whether a robot remained upright for the full test. Survival alone can be misleading: A robot may avoid falling yet make little or no forward progress. The team therefore looked for stable operation paired with a usable gait.

The team also deployed the same policy on real Unitree Go1 and Go2 robots without fine-tuning. Across 10 trials, the Go1 moved indoors and outdoors at about 1.4 to 1.7 meters per second and completed more than 80% of its runs without freezing or suffering a gait failure. It also recovered from pushes. The Go2 reached about 1.1 to 1.3 meters per second.

Hardware tests exposed practical challenges, including control delays, errors in estimating movement, slipping feet and uneven steps. The Go2 also moved more slowly and failed more often under the team’s current deployment setup.

MorFiC has broader limits as well. Transfer depended heavily on how physically similar the target robot was to the training model. The Go2-trained policy worked well on nearby designs such as the Go1, A1 and Cheetah but failed to produce effective movement on the heavier B1 and the more physically dissimilar ANYmal. Simply broadening the range of physical properties during training made the policy more cautious rather than solving the problem.

The researchers suggest training MorFiC on a small number of source robots as a next step—more than one, but far fewer than methods that train across large collections of robot designs. Their current work covers only quadrupeds; extending the framework to two-legged robots and humanoids remains an open challenge.

Watch a video of the robots being tested.

—Story by UMIACS communications group

Research Areas:

Computer Vision and Machine Perception

Ritwik Gupta Joins UMD CS, Studies AI in Complex Environments

Samuel Malede Zewdu — Mon, 21 Sep 2026 14:29:17 +0000

His research focuses on computer vision, active sensing and policy for dual-use artificial intelligence across civilian and defense settings.

September 21, 2026

Artificial intelligence systems often rely on clear images and predictable conditions to understand their surroundings. However, first responders and emergency crews may work amid smoke, damaged infrastructure, poor visibility and incomplete information, all conditions that can limit how well conventional computer vision systems function.

Ritwik Gupta, who joined the University of Maryland’s Department of Computer Science as an assistant professor this fall, studies how to build better AI systems that work with the noisy, unorthodox data first responders commonly encounter. His research brings together computer vision, novel sensing technologies and constraints around the transferable use of AI in both civilian and military settings. Gupta also holds a joint appointment in the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM).

“I am extremely excited to join UMD CS,” Gupta said. “It is one of the rare departments in the world with both meaningful investments and strategy in interdisciplinary research. The faculty and students here are not only incredibly passionate about real-world impact, but also push the state of the art in their respective fields to achieve it. I’m looking forward to contributing to the rich academic culture in the department.”

Much of Gupta’s technical research centers on finding ways for AI systems to gather and interpret information when cameras alone may not be enough. One approach is active sensing, in which a system sends out a signal and measures what returns. Radar uses radio waves, for example, while sonar relies on sound.

“I am focused on representation learning for active sensing modalities such as radar and sonar,” Gupta said.

Representation learning is a subset of machine learning that helps computers identify useful patterns in complex data. Applied to active sensing, it can help an AI system interpret information collected through radar or sonar rather than depending primarily on conventional images.

That capability can become important when AI moves from controlled testing environments into the field. Conditions may change quickly, visibility may be limited and systems may encounter situations that were not represented during development.

Gupta has explored those challenges through projects with real-world applications. His previous research has been deployed to disaster response agencies around the world to assist with search and rescue and reconstruction for hundreds of disasters, helped identify and interdict suspicious activities at sea, and assisted the Ukrainian military in countering the invading forces of Russia.

“My focus is on solving the core theoretical issues that prevent first responders from having access to life-saving AI tools that work in their environments,” Gupta said.

For Gupta, however, deploying AI outside the laboratory also involves questions about how the same technologies may be used in different settings. Sensing and computer vision systems developed for civilian purposes can sometimes rely on methods similar to those used in military applications.

“Further, I focus on dual-use AI policy around the civilian and military use of, often, the same technologies for vastly different purposes,” Gupta said.

Dual-use AI refers to technologies with both civilian and military applications. A sensing system used in disaster response, for example, may share underlying methods with technology used in defense operations. Gupta’s research considers the policy questions that can emerge as similar technologies are developed or deployed for different purposes.

His work in these areas builds on experience spanning academic research, industry and government. Before joining UMD, Gupta led research on AI for humanitarian assistance and disaster response at Carnegie Mellon University and worked on real-time machine learning for Apple Vision Pro. He also served as the Technical Director for Autonomy at the Pentagon’s Defense Innovation Unit.

Gupta earned his doctorate from the University of California, Berkeley, in 2025 and his bachelor’s degree from the University of Pittsburgh in 2017.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Dissolvable Circuit Board Designed for Reuse Earns Design Recognition

Samuel Malede Zewdu — Thu, 17 Sep 2026 15:32:41 +0000

UMD Associate Professor Huaishu Peng and CS alum Zeyu Yan helped develop DissolvPCB, a recyclable circuit board designed to recover materials and electronic components for reuse.

September 17, 2026

Before an electronic device reaches its final form, engineers may build and test multiple versions of its circuit board. Those prototypes can help refine a design, but once they are no longer needed, recovering their materials and electronic components can be difficult.

A research project led in part by the University of Maryland is exploring another approach: building circuit boards that can be dissolved and reused.

DissolvPCB, developed by researchers at UMD, the Georgia Institute of Technology and the University of Notre Dame, received an honorable mention in the student category of Fast Company’s 2026 Innovation by Design Awards, announced Sept. 14. Department of Computer Science Associate Professor Huaishu Peng worked on the project with UMD computer science alum Zeyu Yan (Ph.D. '26, computer science), the study’s first author, and collaborators from the three universities.

The project focuses on printed circuit boards used for rapid prototyping, when engineers build early versions of electronic devices to test and refine their designs. Instead of using conventional circuit board materials, DissolvPCB uses a water-soluble plastic called polyvinyl alcohol, or PVA, as its base and a liquid metal alloy to conduct electricity.

When a board is no longer needed, it can be placed in water. The PVA dissolves, while the liquid metal forms into a bead that can be collected. Electronic components can also be removed and reused. In testing, the researchers reported recovering more than 98% of the materials used to make the boards.

The Fast Company recognition follows a Best Paper Award for the research at the 2025 ACM Symposium on User Interface Software and Technology (UIST). Peng said the latest recognition places the project in a different context.

“This award feels different because it comes from the international design community,” Peng said. “It reflects our goal for this research to break down disciplinary silos and be truly interdisciplinary.”

Beyond reducing discarded materials, the researchers are examining whether reusing circuit board materials and components could lower prototyping costs.

“If we approach sustainability smart enough, it should also generate economic benefits,” Peng said. “With DissolvPCB, the ability to reuse most of the materials and electronic components repeatedly can save circuit designers a large amount of money while also lowering the cost of each design iteration.”

The project brought together Yan, SuHwan Hong and Peng at UMD, Josiah Hester at Georgia Tech and Tingyu Cheng at Notre Dame. Peng said the researchers jointly planned the project and worked through technical challenges as they developed and evaluated the fabrication process.

Yan has since graduated from UMD and looks to join Georgia Tech as a postdoctoral researcher, where he continues research on recyclable electronics.

“Electrification and smartification have been going on for decades, and that will continue with advances in AI and platform technologies,” Yan said. “That also means we’re going to need a lot more electronics and have a lot more e-waste to deal with. We don’t want sustainability to slow that progress down. With such projects, we’re trying to make it practical to manufacture, recover and remake electronics locally.”

The team also developed an open-source software plugin that converts conventional circuit board design files into models that can be produced with a standard 3D printer. Researchers used the method to build and recycle several working devices, including a Bluetooth speaker and a shape-changing gripper.

For now, Peng said DissolvPCB is primarily intended for rapid prototyping.

“In the longer term, it could potentially be extended to the mass production of PCBs if several remaining technical limitations can be overcome,” Peng said.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Human Computer Interaction

Always-On System Unmasks Covert Signals Across the Radio Spectrum

Samuel Malede Zewdu — Mon, 14 Sep 2026 19:14:30 +0000

CS Ph.D. student Aritrik Ghosh and adviser Nirupam Roy developed a low-power system that detects brief covert radio transmissions conventional systems can miss.

September 14, 2026

The SpecSentry prototype uses a custom phase-coded filter and simple detector hardware to identify fleeting transmissions across a wide swath of the radio spectrum while consuming minimal power.

Modern security systems have a dangerous blind spot: a rogue wireless signal can leak sensitive data or disrupt critical infrastructure in less than a millisecond, yet conventional monitors consume too much power to remain constantly vigilant.

A University of Maryland study published earlier this year shows how connected networks can be vulnerable to covert radio transmissions that disappear before standard security systems register them. Traditional wideband monitors require immense power to scan broad ranges of radio frequencies, allowing these fleeting transmissions to escape detection.

To close that gap, the research team developed SpecSentry, a micropower spectrum surveillance system that continuously monitors the airwaves on a tiny energy budget and captures ephemeral wireless threats as they occur. The findings challenge a long-standing trade-off between broad frequency coverage and battery life.

The study was led by Aritrik Ghosh, a fifth-year doctoral student in computer science, with his adviser and study co-author Nirupam Roy, an assistant professor of computer science with an appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS).

The researchers presented their paper, “SpecSentry: Micro-power Wideband Spectrum Surveillance for Ephemeral Transmissions,” in June at the ACM International Conference on Mobile Systems, Applications, and Services (MobiSys) in Cambridge, U.K.

At the heart of SpecSentry is a four-stage “mark-fold-capture-detect” pipeline that enables simultaneous observation across wide frequency bands, demonstrated here over a 400 MHz swath. Unlike conventional systems that scan smaller portions of the spectrum sequentially, SpecSentry captures the entire wideband spectrum at once by deliberately exploiting a normally undesirable characteristic of the filter.

Specifically, a receiver’s passive filter imprints a distinct, frequency-dependent signature on incoming signals. Because the filter draws no power, it adds this marker without increasing the system’s energy demands.

“Filter and antenna engineers have battled for decades with phase distortion in these filters—it’s a defect you normally try to equalize away,” Ghosh said. “We realized this ‘flaw’ could be exactly the signature maker we needed: the filter’s phase response naturally imprints a distinctive, frequency-dependent stamp on any signal passing through it.”

Ghosh compares the process to identifying hidden ingredients in a blended dish. Although blending erases the visual evidence, distinctive spices added to each ingredient remain recognizable to a trained palate. Similarly, SpecSentry uses the filter’s signatures to identify signals hidden within the compressed spectrum.

A simple diode-based circuit then “folds” the broad stretch of spectrum into a compact format, allowing low-power chips to process the data at less than one-twentieth the standard sampling rate. A lightweight AI model untangles the overlapping signatures and determines each signal’s timing, frequency and bandwidth.

When Ghosh and Roy tested SpecSentry across a 400 MHz band, the system achieved near-perfect accuracy—even for transient signals lasting less than a millisecond. Its median error in estimating bandwidth was just 0.3 MHz, a razor-thin margin when monitoring hundreds of megahertz.

Ghosh said covert transmissions might appear for only a few milliseconds before going silent for hours. By continuously monitoring a broad frequency range, SpecSentry can catch them the moment they occur.

Overcoming the conventional trade-off between frequency coverage and energy consumption required rethinking a common assumption in wireless security, Roy said.

“Many people assume that to monitor a wide swath of the wireless spectrum, you need an immense power budget,” Roy said. “Our research shows that by shifting the heavy lifting to smart mathematical folding and lightweight machine learning, we can achieve total vigilance on a microscopic power budget.”

Roy, who is also a core member of the Maryland Cybersecurity Center, said the project has drawn interest from industry and academia as the radio spectrum becomes an increasingly important security domain.

The proliferation of internet of things devices, embedded sensors, autonomous platforms and connected infrastructure has placed billions of wireless devices in homes, cities, factories and military environments. Along with their benefits, these technologies create new opportunities for unauthorized communication, covert wireless activity and spectrum misuse.

Monitoring that activity has become a major security bottleneck in civilian and defense settings, Roy said. By sharply reducing the power and hardware required by traditional systems, SpecSentry could make continuous, widespread monitoring practical.

“This could enable a new generation of spectrum-security sensors that help detect suspicious wireless activity, improve situational awareness, and strengthen the security of increasingly connected environments,” he said.

The technology could have applications in critical infrastructure, smart factories and defense systems that depend on secure, uninterrupted wireless communication.

“Our long-term vision is to make continuous spectrum monitoring practical in places that cannot be covered by a small number of expensive spectrum analyzers,” Ghosh said. “We envision a distributed, always-on wireless security layer that is inexpensive and energy-efficient enough to deploy at scale.”

—Story by Melissa Brachfeld, UMIACS communications group

Research Areas:

Systems, Networking, and Architecture

UMD Students Explore the Mokhtarzada Hatchery Program

Samuel Malede Zewdu — Mon, 14 Sep 2026 19:03:36 +0000

The annual event introduced students to Hatchery’s funding, mentorship and workspace, while offering a chance to hear from its alumni founders.

September 14, 2026

The University of Maryland’s Department of Computer Science held its annual Fall Mokhtarzada Hatchery Kick-Off Event on Sept. 8, giving students an opportunity to learn how the program supports student-led startups from early ideas through product development.

Held at the Brendan Iribe Center for Computer Science and Engineering, the event featured discussions with UMD alumni Haroon Mokhtarzada (B.A. ’01, economics), one of the program’s funders and mentors, and Manpreet Singh (B.S. ’03, finance), a Hatchery mentor. Students learned about the program’s structure, asked questions about the application process and heard how entrepreneurs approach early decisions such as identifying a problem, evaluating a market and developing a product.

The Mokhtarzada Hatchery provides selected teams with up to $10,000 in funding, access to mentors with startup experience and workspace in the Iribe Center. The program supports teams as they develop and test ideas, identify potential users and refine their products and business models.

For some students, the kickoff introduced them to the program.

Meenakshi Sundaram, a freshman computer science major, said the event sparked his interest in the Hatchery.

“I just did a quick Google search, and I saw it was for students who wanted to make a startup,” Sundaram said. “I heard that they were funding it as well with $10,000, so I was like, OK, I’ll just check this out.”

After hearing from the speakers and other students, Sundaram said entrepreneurship became another possible path to consider after college.

“If I was accepted into the program, it would show a whole different path in my life,” Sundaram said. “I was planning on probably just joining some kind of company, but if the Hatchery accepts me, then that would be really cool, and I’d try and put all of my effort into making it a success”

Junior computer science major Neha Joshi came to the event with a different goal. She had previously learned about the Hatchery and attended to better understand how applications are evaluated.

Joshi asked how students could receive feedback on their applications and what the program considers when selecting teams. She said the responses gave her a clearer idea of what to prioritize if she applies.

Joshi is developing an idea for a platform that would allow users to pay influencers or other professionals for short, asynchronous responses to questions. Under the concept, users could submit brief requests and receive voice or video responses without scheduling a call.

“I did get some feedback on my idea,” Joshi said. “That encouraged me more to consider applying to the program.”

The program is funded by UMD alumni and siblings Haroon, Idris (B.S. ’10, computer science) and Zeki Mokhtarzada (B.S. ’01, computer science). The brothers co-founded Webs.com, which VistaPrint acquired for $117.5 million in 2011. They later created the subscription management app Truebill, now Rocket Money, which Rocket Companies acquired for nearly $1.3 billion in 2021.

—Story by Samuel Malede Zewdu, CS Communications

When AI Goes on Defense

Samuel Malede Zewdu — Thu, 10 Sep 2026 15:21:42 +0000

CS Ph.D. student Manan Suri and Professor Dinesh Manocha are part of a research team developing CyberForge, a framework that generates realistic vulnerabilities to train AI agents for software defense.

September 10, 2026

For a cyberattacker, one overlooked weakness can be enough. For a defender, the challenge is broader: find weaknesses and fix them before attackers exploit them.

That imbalance is becoming more consequential as artificial intelligence agents improve at writing and analyzing code. Manan Suri, a University of Maryland computer science doctoral student, is studying the other side of that equation: how AI systems can be trained to repair software vulnerabilities, not just find them.

Advised by Distinguished University Professor Dinesh Manocha, Manan’s recent work on CyberForge, developed through research with the National Institute of Standards and Technology (NIST), Université Grenoble Alpes and CNRS, focuses on one constraint facing defensive AI systems: the amount and quality of cybersecurity training data.

“AI agents are getting stronger at writing code, and as expected, I think that creates a lot of challenges for cybersecurity,” Suri said.

Uneven playing field

AI-assisted coding has increased the speed and scale at which software can be analyzed, raising the stakes for security teams trying to spot weaknesses before attackers exploit them.

“An attacker in a cybersecurity scenario only needs to find one vulnerability to get into your system, whereas a defender ideally needs to find all of them and fix all of them,” Suri said.

AI agents add another dimension to that imbalance. Unlike human operators, they can run continuously, handle multiple tasks at once and scale across large amounts of code. For Suri and his collaborators, that raised a practical research question: whether similar capabilities could be directed toward identifying and repairing weaknesses before they are exploited.

“It’s this asymmetry which creates this race, a race where we must empower defenders as much as we can,” he said.

Data ceiling

Building capable defensive systems depends in part on what they can learn from. In cybersecurity, much of that material comes from records of vulnerabilities that researchers, companies or developers have already discovered and disclosed.

Those examples are useful because they come from actual software, but they represent a finite historical record.

“At one point, you’re going to run out of the historically available data,” Suri said.

The problem becomes more pronounced when a company wants to train a model on its own software. A project may have relatively few documented vulnerabilities, leaving too little material to train an AI system around the specific structure and behavior of that codebase.

“Very often, I would imagine a scenario where an enterprise wants to train a cybersecurity agent that’s specific to their own project,” Suri said. “Individual projects don’t have enough exposed vulnerabilities.”

Synthetic data offers one way to expand that pool, but artificially created weaknesses don't necessarily reflect how vulnerabilities emerge in working software. CyberForge was developed to generate examples using information already contained within a real project.

Building CyberForge

This approach works with open-source C and C++ repositories and looks for places where code can be altered to introduce a security weakness without breaking the software’s normal behavior.

To create a useful training example, CyberForge must introduce a vulnerability without stopping the software from working normally. The researchers can then use a specific input to trigger the weakness and show the model the difference between secure and vulnerable code.

“Given limited historic data, you use the signals from the project to create realistic synthetic vulnerabilities,” Suri said. “What this allows the agent to learn is how to defend itself in case of a real vulnerability.”

The project emerged during Suri’s spring semester as a Professional Research Experience Program fellow at NIST. His background focused on AI agents, while team members brought experience in cybersecurity datasets, vulnerability analysis and standards.

The research brought together researchers from NIST, UMD, Université Grenoble Alpes, and CNRS. In addition to Suri and Manocha, the team included Amine Lbath, Aurelien Delaitre, Vadim Okun, Massih-Reza Amini and Ram D. Sriram.

“I got to work with people who came from a background of cybersecurity,” Suri said. “This is a group of people who have released major datasets and reports and standards for vulnerabilities.”

That combination of expertise also affected how the team evaluated possible research directions. Suri said subject-matter expertise helped distinguish technically plausible ideas from approaches that would be less useful in practice.

“When you’re solving a domain-specific problem, the value that experts bring to that problem is immense,” he said.

Measuring defense

The researchers then tested whether training on CyberForge-generated vulnerabilities made the models better at protecting software. Across several experiments, performance increased by about 3.3 to 15 percentage points, Suri said.

“We wanted to see where open-source models stood on cybersecurity tasks and whether training them on this new data could improve their performance,” he said.

On one benchmark, a model successfully defended against vulnerabilities about 58% of the time before CyberForge training. After training, that figure rose to 72.7%.

The researchers also tested whether the training could help beyond the C and C++ projects CyberForge used. Some models improved on security tasks involving software written in Python, Java and JavaScript.

“We found that the training could carry over to other languages and other cybersecurity tasks,” Suri said.

Manocha said the results point to multiple benefits.

“This approach offers benefits for AI models and cybersecurity,” Manocha said. “It provides higher-quality training signals, scalable generation, broad vulnerability coverage and cross-language generalization. For cybersecurity, it can help produce more capable defense agents and provide coverage beyond known attacks.”

CyberForge addresses the training-data side of defensive AI, but Suri said broader work will also be needed on how security agents are deployed and coordinated across software systems. For defenders, the underlying challenge remains the same: they have to account for more possible weaknesses than an attacker needs to exploit.

“It’s going to be a race,” Suri said. “We should make our agents as well equipped as they can to win.”

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

6 Faculty Members Join UMD’s Department of Computer Science

Samuel Malede Zewdu — Wed, 09 Sep 2026 14:19:47 +0000

New tenure-track and professional-track faculty bring expertise in high-performance computing, internet measurement, computer vision, artificial intelligence, data science and scalable systems.

September 09, 2026

Top row, from left: Yafan Huang, Ritwik Gupta and Ruoshi Liu. Bottom row, from left: Loqman Salamatian, Gheni Abla and Tammy Perrin.

The University of Maryland’s Department of Computer Science is welcoming six faculty members during the 2026-27 academic year, adding research and teaching expertise in high-performance computing, computer networks, computer vision, artificial intelligence, data science and software systems. The appointments include four tenure-track faculty members and two professional-track lecturers joining in fall 2026 and spring 2027.

“We’re thrilled to welcome these new faculty members to the department,” said Department Chair Matthias Zwicker, who holds the Elizabeth Iribe Chair for Innovation and the Phillip H. and Catherine C. Horvitz Professorship. “Their expertise will strengthen research and education across the department, and we are very much looking forward to supporting their work at UMD.”

Yafan Huang

Huang’s research focuses on high-performance computing and systems, with an emphasis on data compression, fault tolerance and parallel computing. His work examines software systems for large-scale scientific computing and artificial intelligence workloads, including GPU-based data reduction, software-directed error resilience and systems support for machine learning.

He has also studied data compression for high-performance scientific computing through collaborations with researchers at Argonne National Laboratory. That work included applications in light source science and seismic imaging involving researchers from Argonne’s Advanced Photon Source and Saudi Aramco’s EXPEC Advanced Research Center.

Huang received the 2025 ACM-IEEE CS George Michael Memorial HPC Fellowship and was named a 2026 MLCommons ML and Systems Rising Star. His work also received best paper awards and finalist recognition at SC 2022, SC 2024, ICS 2025 and LDAV 2025.

He earned his doctorate in computer science from the University of Iowa. Huang joined the department as an assistant professor in fall 2026, with a joint appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS).

Ritwik Gupta

Gupta’s research examines computer vision in complex environments and the policy implications of dual-use artificial intelligence. His work has included applications in post-disaster damage assessment following the 2023 Turkey-Syria earthquake and the detection of illicit maritime activity. He also studies how AI technologies are incorporated into civilian and military systems.

Gupta previously served as technical director for autonomy at the Defense Innovation Unit and held research and policy roles involving AI, security and human rights. Before his doctoral studies, he led an AI-focused humanitarian research lab at Carnegie Mellon University and contributed to real-time machine learning work for Apple Vision Pro.

He earned his doctorate from the University of California, Berkeley. Gupta joined the department as an assistant professor in fall 2026 with a joint appointment in the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM).

Ruoshi Liu

Liu’s research focuses on computer vision systems designed to operate in dynamic and physical environments. His work examines how visual systems can understand and interact with changing real-world settings.

He has developed open-source models and datasets that have been downloaded and reused by researchers and developers. His research has also been discussed in technical communities and public media.

Liu earned his doctorate in computer science from Columbia University.

He joined the department as an assistant professor in fall 2026.

Loqman Salamatian

Salamatian’s research focuses on measuring and understanding the internet despite limited visibility into its decentralized infrastructure. His work develops methods to determine what network properties can be inferred from partial observations and establishes theoretical guarantees for those conclusions.

He applies those methods to study how geopolitical events and other disruptions affect internet connectivity and network structure. His research also examines how network measurements can help diagnose performance problems and support more resilient infrastructure.

His work received a Best of CCR recognition from ACM SIGCOMM Computer Communication Review and was selected as a Communications of the ACM Research Highlight.

He earned his doctorate in computer science from Columbia University. Salamatian will join the department as an assistant professor in spring 2027.

Gheni Abla

Abla's interests span artificial intelligence, software engineering and data science. His work includes building AI and machine learning platforms, developing enterprise data analytics products and creating generative AI applications.

His technical background includes data and AI platforms, GPU-based machine learning acceleration libraries, large-scale data engineering systems for real-time analytics and distributed software designed for supercomputing environments. He has also taught computer science courses in artificial intelligence, machine learning, cloud computing and algorithm design.

Abla earned a master's degree in computer science from the South Dakota School of Mines and Technology and an MBA from the University of Southern California's Marshall School of Business.

Abla will join the department as a lecturer in spring 2027.

Tammy Perrin

Perrin’s teaching and technical interests include artificial intelligence, scalable architectures and high-performance computing. She has taught courses in agentic artificial intelligence, big data systems and data science.

Perrin has more than 15 years of technology industry experience, including work at Google, Capital One and Lockheed Martin.

She earned a master’s degree in computer science from Johns Hopkins University and a bachelor’s degree in physical sciences from the University of Maryland, with studies in electrical engineering, mathematics and physics. Perrin is a two-time recipient of the DCFemTech Powerful Women Programmers Award.

Perrin joined the department as a lecturer in fall 2026 and teaches in the University of Maryland Science Academy’s graduate programs.

—Story by Samuel Malede Zewdu, CS Communications

In Pictures: 2026 Welcome Back Social

Samuel Malede Zewdu — Tue, 08 Sep 2026 17:45:20 +0000

The annual fall event brought students, faculty and staff together at the Iribe Center to mark the start of the fall semester.

September 08, 2026

Students, faculty and staff gathered at the Brendan Iribe Center for Computer Science and Engineering on August 31 for the University of Maryland Department of Computer Science’s annual Fall Welcome Back Social.

Held on the first day of classes, the social gave new and returning students an opportunity to meet classmates, connect with department representatives and student organizations, and learn about resources available throughout the academic year. The department begins the fall semester with about 3,300 undergraduate students and 612 graduate students.

Previous

Next

—Portrait Article by Samuel Malede Zewdu, CS Communications

NSF Renews Maryland-Led Quantum Simulation Institute’s Funding

Samuel Malede Zewdu — Thu, 03 Sep 2026 14:40:19 +0000

A five-year NSF award expected to total $37.5 million will support the next phase of the Maryland-led Institute for Robust Quantum Simulation.

September 03, 2026

A University of Maryland-led institute focused on robust quantum simulation will launch a new phase of research with a five-year U.S. National Science Foundation award expected to total $37.5 million.

The renewal for the NSF Quantum Leap Challenge Institute for Robust Quantum Simulation (NSF RQS), which takes effect Sept. 1, reflects UMD’s long-term commitment to advancing quantum computing to take on society’s grand challenges and to spark a new era of scientific, technological and economic development, UMD President Darryll J. Pines said.

“Our researchers in quantum simulation are an important part of the vibrant quantum ecosystem we’ve built,” Pines said, “helping establish our campus and the surrounding region as the Capital of Quantum and making our Discovery District a dynamic hub where research, industry and community combine for real-world impact on the public good.”

Established in 2021 with a $25 million federal award, NSF RQS has developed novel ways to verify quantum systems, reduce errors and demonstrate increasingly sophisticated quantum simulations while building a collaborative community of scientists, engineers, educators and students.

Quantum simulation is widely viewed as one of the first practical applications of quantum computing. Rather than trying to make every type of computation faster, quantum simulators are designed to model extraordinarily complex quantum systems that overwhelm even today’s most powerful conventional computers.

The new award expands the institute’s scientific agenda; sustains education and workforce development programs; strengthens partnerships across academia, federal laboratories and industry; and positions NSF RQS to pursue the next generation of challenges in quantum simulation. Harvard University also will join the NSF RQS consortium, expanding the current collaboration between UMD, Duke University, Princeton University, Yale University and researchers from the National Institute of Standards and Technology (NIST).

NSF RQS is one of eight NSF Quantum Leap Challenge Institutes (QLCI), a network of interdisciplinary research centers created to accelerate advances in quantum information science through collaborative research, education and workforce development.

“For more than four decades, NSF has been laying the foundational groundwork of research and discovery that is powering today’s modern quantum computing, sensing and communication,” said Brian Stone, performing the duties of the NSF director. “It’s time for focused activities to leverage that base of knowledge to drive us even farther forward to the benefit of all Americans. The NSF Quantum Leap Challenge Institutes are a next step for us in understanding the quantum world we live in.”

During its first funding cycle, NSF RQS researchers achieved breakthroughs in fault-tolerant quantum simulation with neutral atoms, new methods for verifying quantum advantage on analog quantum simulators, erasure detection in Rydberg atoms and quantum simulations of gauge theories and quantum materials. Together, these advances brought robust quantum simulation closer to becoming a practical tool for scientific discovery.

That progress was matched by the institute’s growth as a national research enterprise. NSF RQS researchers collectively produced more than 600 papers, generating nearly 17,000 citations. Along the way, NSF RQS trained more than 400 graduate students and postdoctoral researchers while developing programs designed to strengthen the nation’s future quantum workforce.

“NSF’s investment in the QLCIs established hubs for research on critical topics in quantum information science at a time when the field is rapidly advancing,” said Andrew Childs, a UMD professor of computer science who directed NSF RQS during its first five years. “I’m grateful for all the collaborations it enabled and proud of the progress RQS researchers made in just five short years.”

Childs is stepping down as director after guiding the institute through its formative years. Mohammad Hafezi, a UMD Minta Martin Professor with joint appointments in physics and electrical and computer engineering, will become director. Michael Gullans, a physicist at NIST and an adjunct assistant professor in physics and the University of Maryland Institute for Advanced Computer Studies (UMIACS), will serve as deputy director.

Mohammad Hafezi, a Minta Martin Professor of Physics and Electrical and Computer Engineering at the University of Maryland, will become director of the NSF Quantum Leap Challenge Institute for Robust Quantum Simulation under its renewed five-year award. Image courtesy of NSF RQS

“We are at the cusp of determining exactly what quantum computers can and cannot do in the near future,” Hafezi said. “Realizing that potential will require vibrant partnerships among universities, federal laboratories and private industry, and we’re grateful for this renewed support to help lead that effort.”

The new award funds a shift toward what institute leaders call “quantum simulation engineering”—developing quantum technologies that are increasingly scalable and capable of addressing important scientific problems. Researchers will focus on three research themes: interacting fermion simulation, which seeks to model the particles that make up matter; fermion-boson simulation, which examines how matter particles interact with force-carrying particles in complex quantum systems; and dissipative quantum simulation, which studies quantum systems interacting with their environments instead of treating environmental effects solely as unwanted noise.

NSF RQS benefits from technical and administrative support provided by UMIACS, which helps coordinate research activities across the institute’s university and federal partners, Hafezi said.

The institute will also expand its educational mission alongside its research portfolio.

During its first funding cycle, NSF RQS helped launch the University of Maryland’s quantum science and engineering minor, supported K-12 teacher professional development workshops that reached 75 educators, and developed innovative quantum activity toolkits used by 100 teachers to introduce more than 2,000 middle and high school students to concepts like superposition and measurement.

The institute also established an international QSim conference series that attracted more than 600 attendees and launched the Quantum Leap Career Nexus, which connects students with more than 50 companies, government laboratories and universities through career fairs and networking opportunities.

The renewed award will build on those efforts through additional teacher development workshops, public outreach, K-12 quantum education programs, research seminars and summer schools while creating new opportunities for students and postdoctoral researchers to prepare for careers throughout the quantum workforce.

Gretchen Campbell, associate vice president for quantum research and education at UMD, will lead the institute’s education and workforce development efforts moving forward.

“Preparing the future quantum workforce means creating opportunities at every stage—from K-12 classrooms to graduate education and professional careers,” Campbell said. “By connecting students, educators and researchers across that entire pipeline, we can help ensure the talent needed to advance quantum science and technology is ready when the next breakthroughs arrive.”

—Story by UMIACS communications group

Research Areas:

Quantum Computing

UMD Course Connects Computer Science, Math and Sports Analytics

Samuel Malede Zewdu — Tue, 01 Sep 2026 15:16:57 +0000

MATH 440: “Geometry and Data Science of Sports Analytics” gives students experience applying programming, data analysis and machine learning to careers in professional sports.

September 01, 2026

Credit: Adobe Stock Images.

Growing up, Jason Lott (B.S. ’25, computer science and mathematics) was always fascinated by sports. So, when he was scrolling through Testudo during his junior year at the University of Maryland to sign up for classes, he was immediately attracted to MATH 498R: “Experiential Learning: Mathematics of Sports Performance Analytics,” designed and taught by Mathematics Professor Yanir Rubinstein.

Lott needed a math elective to fulfill his major requirements, and with his lifelong passion for major league baseball, the class sounded like a perfect fit.

“It didn’t seem like the usual math class,” he recalled. “It was about sports analytics, and I’ve always been interested in athletes’ stats. I felt like I had to take it. And obviously, it’s worked out for me, considering where I ended up after graduation.”

Today, Lott works as a junior software engineer in the Baltimore Orioles' baseball analytics department. He helps maintain the systems that house the club's data and supports an internal website and mobile app used by coaches, scouts, analysts and players to improve their performance. Several types of data, from players’ batting averages to sprint speed, become the foundation for important decisions made by the team’s management, like training optimization or roster development.

“We store data about everything you could imagine—pretty much any stat you can think of surrounding the sport of baseball, including historical records going back many years,” Lott explained. “All those stats are used by analysts trying to gain any type of edge on the field or in the front office.”

Lott credits the experiential math course for helping him navigate his career path toward the Orioles and a professional sports job he always wanted. He said the class, which tied together his background in both computer science and math, allowed him to directly explore complex, real-life scenarios in sports—something that no other college course had ever done.

A math course unlike any other

First piloted in spring 2023, the UMD sports analytics course grew out of Rubinstein’s UMD Grand Challenges Award and his broader effort to connect students and the wider community to STEM through sports. After two successful pilot runs, the course was approved as a permanent offering for all undergraduate students. Beginning fall 2026, it will be offered as MATH 440: “Geometry and Data Science of Sports Analytics.”

For Rubinstein, the new class marks a milestone for the Department of Mathematics.

“We’ve built a data science curriculum that is from a mathematical perspective,” Rubinstein said. “With the training our students receive here, they are able to open more doors after they graduate.”

MATH 440 is structured less like a traditional math class and more like an industry internship. Instead of homework, quizzes and exams, students work in teams on semester-long research projects that simulate what it’s like to work for a professional sports team. Rather than being told which model or strategy to use, students experiment and figure out what works for themselves. By the end of the course, students learn how to wrangle real data and present findings in a clear, understandable way just like it’s done in a real analytics job.

In 2024, when Lott took the analytics class, the assigned sport was basketball. He and his partner were tasked with figuring out if machine learning could reverse-engineer the formulas behind popular basketball statistics.

“We had one simple stat where you can look up the formula, so we could compare our results against the real thing,” Lott explained. “The harder challenge was taking a black-box formula—a proprietary stat where the creator keeps the calculation secret—and trying to learn how it was calculated.”

The course also introduced students to using APIs (a programming tool that allows developers, researchers and fans to automatically extract and integrate NBA data). Lott and his partner learned to pull live data, including individual player statistics and historical team records, from the NBA's API for their project.

“Knowing how to extract data from an API and save it is a big part of my job now,” Lott said. “I use the skills I developed from the class at work every day, and I’ll probably continue to for years to come.”

For Rubinstein, the course is more than getting math students into the major leagues. Seeing students like Lott thrive in the sports industry reflects a bigger goal: feeding young people’s interest in math by showing them that it has applications they can use in the real world. Students from the course also present their projects in K-12 schools across the D.C., Maryland and Virginia, as part of Rubinstein’s Mobile Mathematics Motion Lab. Rubinstein’s ultimate goal is to combine teaching, research and community engagement in a way that changes perceptions of what mathematics—and a university mathematics department—can do.

“Oftentimes, students lose their interest or passion for math at young ages because they have misconceptions about what math is and where they can use it,” Rubinstein said. “What my team is doing isn’t really outreach or mathematics education in the usual sense. We’re teaching things that we’re not usually teaching because we want to connect mathematics to people’s real lives.”

—Article adapted from a story by CMNS

UMD CS Organizes 2026 Research Day

Samuel Malede Zewdu — Mon, 31 Aug 2026 14:02:48 +0000

Five panels introduced the department’s 179 incoming graduate students to research in theory, artificial intelligence, robotics, systems and cybersecurity.

August 31, 2026

The University of Maryland’s Department of Computer Science organized its annual Computer and Computational Sciences Research Day on August 27, giving incoming graduate students an overview of research across computing fields and opportunities to connect with faculty. The event took place in Antonov Auditorium at the Brendan Iribe Center for Computer Science and Engineering.

The department expects 179 incoming graduate students this fall. The cohort includes 78 doctoral students, 49 master’s students and 52 students pursuing a combined bachelor’s and master’s degree.

Faculty presentations were organized into five panels: “Foundations: Theory, Language, Quantum & Parallelism”; “AI for Science, Society & Real-World Impact”; “Machine Perception & Sensing”; “Robotics, Autonomy & Embodied Intelligence”; and “Systems, Networking & Cybersecurity.” Breaks between sessions gave students time to discuss the presentations with faculty members and one another.

The event featured research from the Department of Computer Science and several UMD institutes, centers and laboratories, including the University of Maryland Institute for Advanced Computer Studies (UMIACS), the Maryland Cybersecurity Center (MC2), the Human-Computer Interaction Laboratory and the Center for Bioinformatics and Computational Biology.

Amol Deshpande, a computer science professor and the department’s associate chair for graduate education, described how the format introduces students to several subjects.

“Research Day gives incoming graduate students an opportunity to learn about work taking place across the department,” Deshpande said. “For those who have not already narrowed their interests, it is a chance to meet faculty members and other students, ask questions, and begin identifying potential advisers or research groups.”

David Mazi, an incoming master’s student planning to study artificial intelligence and machine learning in health care, said the foundations panel prompted him to consider AI’s computing infrastructure. He previously studied electrical and computer engineering at the University of Arizona.

Mazi said Affiliate Professor of Computer Science Uzi Vishkin’s presentation introduced him to parallel computing in a simpler way than he had previously examined.

“I would like to not just understand what I want to focus on, but understand every kind of adjacent piece of technology around it,” Mazi said. “Knowing about hardware greatly informs the kind of software you should be writing to take advantage of that.”

The AI panel also connected with Mazi’s interest in health care. He cited Assistant Professor Ritwik Gupta’s discussion of disaster analysis and another presentation on using AI in education as possible connections outside his primary area.

Nishit Anand entered UMD’s doctoral program after completing his master’s degree in computer science at the university in 2026. He worked with professors Dinesh Manocha and Ramani Duraiswami on benchmarking and training large audio-language models that process speech, music and background sounds.

Anand said the range of subjects was the main feature that stood out to him during Research Day.

“It was not limited to just one area,” Anand said.

He pointed to presentations on computer vision, robotics, natural language processing, audio processing, dexterous manipulation and diffusion models. Anand said he plans to examine diffusion models, audio understanding and audio processing while determining a specific doctoral research direction.

The department’s graduate program is No. 12 overall and No. 7 among public institutions in U.S. News & World Report’s 2027 Best Graduate Schools rankings. Artificial intelligence ranks No. 16, systems No. 17 and theory No. 19.

—Story by Samuel Malede Zewdu, CS Communications

Targeted Ads Come with Unavoidable Privacy Trade-Offs, UMD Study Finds

Samuel Malede Zewdu — Fri, 28 Aug 2026 17:56:53 +0000

Research co-authored by Gabe Kaptchuk found that combining ad targeting with campaign performance data can expose private user information despite technical safeguards.

August 28, 2026

Tech companies and browser developers have spent years pursuing a seemingly ideal solution to one of the web’s thorniest problems: using advanced cryptography to make online advertising useful without sacrificing user privacy.

New research co-authored at the University of Maryland shows there is a fundamental limit to that goal. Perfect privacy and useful targeted advertising, the researchers found, simply cannot coexist.

The findings, presented last month at the Privacy Enhancing Technologies Symposium (PETS) in Canada, challenge a foundational hope in privacy engineering. Rather than treating ad targeting and campaign measurement as separate features, the researchers zoomed out to examine them as two sides of the same coin. In doing so, they revealed that trying to protect user privacy while running both mechanisms simultaneously creates systemic vulnerabilities—a big-picture conflict that previous studies had overlooked by analyzing each piece in isolation.

Gabe Kaptchuk, a study co-author and UMD assistant professor of computer science, said the work demonstrates why the security of an advertising system cannot be evaluated piece by piece.

“Our research shows that when you connect targeting and performance feedback, the interaction introduces a fundamental, unavoidable layer of information leakage,” he said.

Along with Kaptchuk, the research team includes lead author Kyle Hogan, an electrical engineering and computer science Ph.D. student at the Massachusetts Institute of Technology (MIT); Alishah Chator, assistant professor of mathematics at Baruch College; Mayank Varia, associate professor of computing and data sciences at Boston University; and Srinivas Devadas, the Edwin Sibley Webster Professor of electrical engineering and computer science at MIT.

Kaptchuk, who has an appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS), said the study grew out of efforts across academia and the technology industry to rethink online advertising amid mounting concern over the extensive tracking that underpins much of the digital ad ecosystem.

“Digital advertising is incredibly invasive, as it relies on gathering a tremendous amount of information about individuals,” he said.

That tracking can cause concrete harms. Commercial ad networks can build profiles around sensitive characteristics—identifying, for example, people believed to be “depression-prone” or “credit-reliant.” Such profiling has been shown to enable predatory marketing, discriminatory job or housing advertising, and manipulative political microtargeting.

But the researchers found that efforts to make advertising more private have overlooked a basic contradiction: The performance information advertisers need to evaluate and refine campaigns can itself reveal information about users.

Modern advertising systems connect two critical functions—deciding who receives an ad and reporting how that ad performs. Even if each function incorporates privacy protections, combining them can create new vulnerabilities.

An advertiser, for example, could repeatedly target carefully selected groups, such as people on a customer list, and analyze the resulting performance metrics. Over time, those results can allow the advertiser to statistically infer private information about users—even when cryptographic protections prevent their identities from being directly revealed.

For browser vendors and ad-tech companies developing the next generation of web advertising standards, the findings suggest that adding privacy protections to individual components will not be enough.

“Simply throwing privacy-enhancing technologies at the problem won't work—we need to be careful about designing interventions,” said Kaptchuk, who is also a core faculty member in the Maryland Cybersecurity Center. “If we’re unable to make advertising private on a technical level, then we require robust social mechanisms to ensure that advertisers use digital advertising platforms responsibly.”

Rather than pursuing complete technical secrecy, the researchers propose a design framework that accounts for how targeting choices interact with performance data. Future advertising systems, they argue, should be “targeting-aware,” applying group-level protections that limit what advertisers can learn about sensitive attributes such as health, finances or political affiliation.

The findings point toward a broader approach to online privacy—one that combines carefully designed technical protections with rules governing how advertising platforms and advertisers can use the information available to them.

For Kaptchuk, that means recognizing the limits of what cryptography alone can accomplish and designing protections around those limits.

—Story by Melissa Brachfeld, UMIACS communications group

Research Areas:

Inside TechStart 2026

Samuel Malede Zewdu — Wed, 26 Aug 2026 15:50:21 +0000

Computing Catalyst’s TechStart program gave University of Maryland students a week to explore coding, design and technology through community-focused projects.

August 26, 2026

TechStart participants come together for a final group photo after a week of coding, collaboration and project development. Credit: Samuel Malede Zewdu

A tool to help students make connections on campus. A redesigned University of Maryland map. A way to find cooler spaces during hot summer days.

Those were among the projects UMD students presented as the Computing Catalyst’s 2026 TechStart program concluded. Across separate virtual and in-person sessions, participants spent a week identifying problems, considering potential users and creating technology-based solutions.

TechStart is a paid program for students interested in exploring computing, including those with little or no previous coding experience. This year’s in-person and virtual sessions included students studying computer science, information science, government and politics, and other fields, as well as students still considering a computing major.

Participants took part in programming, design thinking activities and career discussions. They also met with technology professionals who shared perspectives on careers and industry work.

“While TechStart teaches students Python and app development, building community is just as central to the program,” said Kristina Kramarczuk, associate director of the Computing Catalyst. “Incoming students connect with peers and industry mentors, creating a supportive network before their first semester even begins.”

Computer Science lecturer Elias Gonzalez led sessions on Python data structures, artificial intelligence and prototyping platforms, introducing students to tools and concepts they could use throughout the program.

“TechStart is one of my favorite things to teach,” said Gonzalez, the program’s lead lecturer. “This year, the students did a great job with their industry mentors to solve tech-for-good problems. I’m excited to see how they continue to grow at UMD and with Computing Catalyst.”

The students’ work took different forms as teams moved from early ideas to project prototypes.

Sophomore technology and information design major Kamryn Coleman worked on an application concept focused on media literacy in the workforce. Her team examined scams involving manipulated voices or likenesses and considered how a digital tool could help users recognize potentially deceptive content.

“Working on this project taught me that collaboration is key to forming good ideas,” Coleman said.

As the project developed, Coleman also took on additional responsibility for moving it through its remaining stages and refining the design with support from mentors.

“Being able to adapt to any situation was also key to succeeding,” Coleman said.

Other teams faced different design decisions.

Freshman information science major Rosy Angwi joined TechStart with no prior coding experience. Her team developed an application concept that would synchronize calendars across educational platforms, bringing deadlines, notifications and academic resources into one place.

“I saw TechStart as an opportunity to bridge the knowledge gap,” Angwi said. “Additionally, I wanted to meet my professors and connect with potential classmates.”

By the end of the program, Angwi said the experience had changed some of her assumptions about working with technology.

“The program has transformed my approach to tech,” Angwi said. “It has taught me that it demands far more than mere code copying.”

Program staff members saw similar changes as students moved through the week.

Anusha Saha, a TechStart program assistant and junior computer engineering major, worked with participants during classroom activities, mentor sessions and project development.

“I have noticed that the students became more confident both in their knowledge and in their interactions with their peers,” Saha said. “They were more comfortable answering questions in class and participating in discussions, and they grew more confident in reaching out to and working with their mentors. That’s what TechStart is all about.”

—Story by Samuel Malede Zewdu, CS Communications

Can AI See the Next Flood Coming? UMD Team Aims to Transform Water Forecasting

Samuel Malede Zewdu — Mon, 24 Aug 2026 18:20:34 +0000

A UMD-led team, including Chris Metzler, Abhinav Bhatele and Haizhao Yang, is developing an AI framework to produce faster, more detailed long-range water forecasts.

August 24, 2026

Predicting whether next season will bring a devastating flood or a crippling drought often comes down to a race against the clock. For climate scientists, forecasting the future movement of water requires some of the most complex—and computationally demanding—math on Earth.

Across the United States, water-related hazards threaten electrical grids, reservoirs and other critical infrastructure. Energy providers need localized, long-range forecasts to prepare for these risks, but current prediction systems are limited by scientific uncertainty and the enormous computing power required to run them.

Now, a University of Maryland-led team is using generative artificial intelligence to change that.

The researchers have been awarded $750,000 in Phase I funding from the U.S. Department of Energy's new Genesis Mission to develop an AI-powered framework that can generate faster, more detailed seasonal and multi-year water forecasts. The nine-month award includes approximately $470,000 for UMD, the lead institution.

The project, "An Agentic AI Framework for Seasonal-to-Interannual U.S. Water Prediction," was announced last month at the inaugural Genesis Mission Summit in Washington, D.C. It was one of 270 projects selected from more than 5,000 proposals submitted to the highly competitive national program. If chosen for Phase II, the team will expand its early prototypes into operational tools with broader real-world applications.

Led by Jinwoong Yoo, an assistant research scientist at UMD's Earth System Science Interdisciplinary Center, the project brings together experts in atmospheric science, artificial intelligence, high-performance computing and mathematics.

The UMD team includes four faculty members from the University of Maryland Institute for Advanced Computer Studies (UMIACS): Maria Molina, assistant professor of atmospheric and oceanic science; Chris Metzler, assistant professor of computer science; Abhinav Bhatele, professor of computer science; and Haizhao Yang, professor of mathematics and Affiliate Professor of computer science.

Researchers from Oak Ridge National Laboratory, the University of Florida, the University of Colorado Boulder and Silurian AI are also collaborating on the effort.

For Molina, the partnership reflects a growing need to combine advances in AI with decades of expertise in Earth system science. The team's goal is to integrate cutting-edge AI forecasting with the Department of Energy's flagship Energy Exascale Earth System Model (E3SM), creating a system that can produce faster predictions without sacrificing scientific rigor.

“If we can get this framework right, we can give planners a clearer picture of the hazards their communities may face seasons to years in advance by revealing how event likelihood, intensity and local impacts may evolve,” Molina said.

Traditional physics-based Earth system models such as E3SM simulate interactions among the atmosphere, ocean, land and water cycle across the globe. Seasonal-to-multiyear forecasting requires many simulations to capture the range of possible outcomes, creating tradeoffs among resolution, forecast length, and ensemble size. Turning those forecasts into useful information at the community level remains difficult because predictability decreases at longer timescales, while local hazards depend on processes that global models cannot fully represent.

The UMD team's approach uses AI to accelerate nearly every step of that process.

The framework will first incorporate satellite observations of soil moisture and groundwater using machine learning to improve the model's starting conditions. Generative AI models trained on decades of E3SM simulations will then emulate the behavior of the original model, producing more than 100 possible forecast scenarios to estimate both likely outcomes and uncertainty more than 100 times faster than traditional methods.

Image-enhancement techniques will sharpen those regional forecasts into street-level flood maps with one-meter resolution. Behind the scenes, autonomous AI software agents will coordinate the complex workflow—from preparing model inputs to running simulations and analyzing results—allowing researchers to focus on evaluating forecast quality and scientific reliability.

To ensure the predictions remain grounded in physical science, Molina will lead the team's validation efforts. Researchers will test the framework against major hurricanes, evaluate whether it captures climate patterns such as El Niño and use explainable AI techniques to determine which factors drive the model's forecasts.

"The real bottleneck has always been computational cost," Metzler said. "By using AI to emulate these massive Earth system models, we can run simulations in minutes on supercomputers that used to take days. That speed is what makes real-time protection of local grids actually possible."

Before deploying the system on the Department of Energy's leadership-class supercomputers, the team will develop, train and evaluate its initial models using UMIACS' high-performance computing infrastructure.

Ultimately, the researchers hope their work will help utilities, water managers and emergency planners make better decisions before disasters strike.

For the team, success won't be measured simply by faster algorithms or more efficient computing. It will be measured by more accurate forecasts that help communities prepare for floods, protect critical infrastructure and better manage water resources before extreme conditions arrive.

—Story by Melissa Brachfeld, UMIACS communications group

Research Areas:

Natural Language Processing

The Code They Couldn't Talk About

Samuel Malede Zewdu — Fri, 21 Aug 2026 19:07:10 +0000

UMD alum Eileen Buckholtz spent three decades working with emerging technology at the NSA. Decades later, she is documenting the computing careers of women whose classified work kept them largely out of history.

August 21, 2026

For much of the career of alum Eileen Buckholtz (M.S. ’73, computer science), the details of her work at the National Security Agency remained private. She could discuss projects with colleagues, but not with her husband, children or friends. Some accomplishments could not appear on a resume.

That secrecy also meant the experiences of Buckholtz and other women working with computers at the agency were largely absent from the public record.

More than five decades after beginning her career, Buckholtz is revisiting that history through “Queens of Code,” a project documenting the women who worked in computing at the NSA.

“There were more than 150 of us, but I wrote about 25 in my book,” Buckholtz said. “We were almost all hired as 21- and 22-year-old women, and we began to program and manage the most sophisticated computers in the world.”

Early computing

Buckholtz entered the field when computer science was still developing as an academic discipline. After earning an undergraduate degree in computer science from Ohio State University, she joined the NSA at 21.

She later earned her master’s degree in computer science from the University of Maryland in 1973, the same year UMD established its Department of Computer Science.

At Maryland, Buckholtz encountered technologies and ideas that were reshaping computing. A discussion in a data communications course introduced her to ARPANET, the early network that preceded the modern internet. At NSA, she connected to the network for the first time and sent a message to a researcher at the Stanford Research Institute, now SRI International, in California.

“There were a lot of possibilities,” she said.

Her master’s research explored potential computer applications in a cable television environment, including concepts such as online education and shopping. Her work, advised by Professor Emeritus Victor Basili, was later published in Communications of the ACM, and she presented her research at the 1973 ACM Annual Conference.

Buckholtz also carried lessons from UMD back to the NSA. After taking a course on formal requirements for projects, she brought the methodology to the agency, where she said it was later adopted for requirements specification.

A Career Behind Secrecy

During her 31-years at the NSA, Buckholtz worked across a range of emerging technologies and computer systems.

She served as the first project manager for a multimillion-dollar UNIVAC 1110 system and worked in areas including office automation, word processing, CD-ROM development, library automation and machine translation. She also helped develop the agency’s first five-year plan for computers and telecommunications.

Much of that work could not be discussed outside the agency.

“It’s hard because you couldn’t tell your husband or your kids what you were doing,” Buckholtz said. “For some of the women who tell their stories now, and my family too, it was the first time they had heard about the things that we had done.”

The restrictions went beyond conversations at home. Classified projects couldn't be publicly described or included on resumes.

Decades later, that experience would help motivate Buckholtz to document the careers of the women who had worked alongside her.

Finding Missing Stories

Buckholtz had long wanted to write about her NSA experience, but she understood the challenges of publishing material involving classified work.

Then a visit to the Computer History Museum in California gave the idea renewed urgency.

As she walked through the exhibits, Buckholtz noticed a history dominated by machines and male figures, with relatively few women represented.

“That was not the environment that I knew,” she said.

In 2018, Buckholtz began working with NSA historians and the National Cryptologic Museum to determine how to preserve those experiences while complying with the agency’s prepublication review process. The process differed from a typical oral history. Former colleagues first had to write their own accounts and receive NSA approval before Buckholtz could review them.

History revisited

Reconnecting with former colleagues changed Buckholtz’s understanding of the work taking place around her at the NSA. Because employees often knew only part of what others were doing, she said she had not fully understood the scope of some colleagues’ contributions until she began researching their careers.

That work led her to projects in areas including microelectronics and cybersecurity, while also revealing how often mentorship shaped those careers.

Buckholtz said she worked for nine women during her NSA career, many of whom helped guide her professionally. She later saw that pattern continue as women who had benefited from mentorship early in their careers supported the generations that followed.

Among Buckholtz’s own mentors was Dottie Bloom, an NSA supervisor who gave her opportunities to take on technical and planning projects.

Looking back through those experiences has also changed how Buckholtz sees her own place in computing history.

“I didn’t when I first started, but I guess I see myself now as a historian,” she said.

Her research has since expanded beyond individual biographies to examine the broader history of computing at the NSA. For Buckholtz, documenting that history is also a way to show current students who participated in the field’s early development.

“I think it’s really important to women in STEM today to be able to know that there were women who were here at the beginning of computing,” she said. “I hope that the work that I do of bringing these stories to light helps them feel like they belong and that there’s a future for them.”

—Story by Samuel Malede Zewdu, CS Communications

###

Queens of Code: Inspiring True Stories from NSA’s Computing Women will be published by MIT Press on Nov. 17, 2026.

Other UMD graduates featured in the book include:

Carol Osborne McWilliams (B.A. ’67, english)

Peggy Ann Grindrod Strader (B.A. ’66, education)

Kathy Reading (B.S. ’98, management)

Mary T. Clulow (B.S. ’97, information systems management)

Readers, including students interested in learning more about Buckholtz’s work, can connect with her through the Queens of Code website, LinkedIn or the Queens of Code Facebook page.

CS Researchers Receive AIM Seed Grant to Study the Roots of AI Bias

Samuel Malede Zewdu — Tue, 18 Aug 2026 19:13:24 +0000

Naomi Feldman, Rachel Rudinger and Sarah Wiegreffe will examine how cognitive biases influence language model decisions and test methods for reducing them.

August 18, 2026

A team of University of Maryland Department of Computer Science researchers has received seed funding for an interdisciplinary effort that combines cognitive psychology and computer science to investigate how artificial intelligence makes decisions—and why those decisions can become irrational or biased.

Supported by a one-year grant totaling just under $120,000—half from the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM), matched by funding from the researchers’ academic units—the project brings together three faculty members with appointments in the University of Maryland Institute for Advanced Computer Studies (UMIACS).

The collaborators are Naomi Feldman, professor of linguistics; Rachel Rudinger, associate professor of computer science; and Sarah Wiegreffe, assistant professor of computer science. All three are members of UMD’s Computational Linguistics and Information Processing (CLIP) Lab, which is supported by UMIACS.

The CLIP researchers are tackling a fundamental challenge in modern AI. Although large language models are increasingly used to support decisions in areas such as hiring and college admissions, they can be influenced by information that should have no bearing on the outcome. In human psychology, these errors are known as cognitive biases—for example, the anchoring effect, in which an unrelated number unconsciously influences a person's estimate or judgment. The UMD team has found that language models exhibit many of these same irrational tendencies.

Feldman's lab brings extensive expertise in computational cognitive science—using computer models to understand human thought processes—and is now reversing that approach by applying theories of human cognition to better understand machine behavior.

“We’ve known for a long time that people reason irrationally,” said Feldman. “Anytime people need to estimate a quantity or proportion to decide, they are subject to cognitive biases. In any decision-making context involving these models, we want to make sure decisions are based on facts in a rational manner rather than on irrelevant information in the surrounding context.”

While these cognitive biases may seem relatively harmless, the researchers believe they could also create pathways for harmful social biases, such as racial or gender stereotyping, to influence AI-assisted evaluations like resume screening and admissions decisions.

“Cognitive bias and social bias are often studied separately, but we think they're deeply connected,” said Rudinger. “If we can understand why language models rely on irrelevant information when making decisions, we'll be in a much better position to reduce the unfair biases that can emerge in real-world applications.”

To move from observation to intervention, Wiegreffe’s lab focuses on mechanistic interpretability—the study of the internal circuits that drive a model's behavior. By identifying the specific components that respond to irrelevant context or demographic attributes, the team hopes to eventually disable, or “ablate,” those circuits without diminishing the model's overall capabilities.

“The key challenge is that models are undesirably influenced by context, but this influence is inconsistent,” said Wiegreffe. “When AI systems are used in high-stakes settings like job interviews and resume reviews, our goal is to give people tools to control those models. We want users to be able to ablate specific circuit components so they can trust that the model is truly ignoring race or gender in a predictive decision.”

Feldman said the collaboration grew out of work by CLIP graduate students. Her Ph.D. student, Hillary Owusu, began investigating AI cognitive biases for a class project and discovered strong connections with research on AI and human reasoning led by Rudinger and her Ph.D. student, Shramay Palta. The AIM grant will provide full support for both senior doctoral students over the coming year.

The work also relies heavily on UMIACS’ advanced computing infrastructure. Because analyzing the internal representations of large language models requires significant computational power, the researchers use the CLIP and UMD Center for Machine Learning GPU clusters at UMIACS to train, run and analyze large open-source models, providing access to internal model data that commercial AI platforms do not make available.

Over the next year, the team plans to expand its dataset of cognitive testing benchmarks, evaluate new intervention strategies and use the results to pursue a larger collaborative proposal to the National Science Foundation. Ultimately, the researchers hope their work will not only explain why AI systems make biased decisions but also provide practical tools that help people identify and reduce those biases before AI is used to make consequential decisions.

—Story by Diya Sharma, UMIACS communications group

Research Areas:

Noriko Maitland Receives UMD Provost’s Academic Advisor of the Year Award

Samuel Malede Zewdu — Tue, 18 Aug 2026 14:30:40 +0000

She was selected for the 2025-26 academic year for her work on advising resources, student communication and departmental initiatives.

August 18, 2026

University of Maryland Department of Computer Science Academic Advisor Noriko Maitland has received the Provost’s Academic Advisor of the Year Award for the 2025-26 academic year.

Each year, a committee of academic advisors selects three university employees for the award, which recognizes advising contributions across campus. Maitland will receive a plaque and a $1,000 award and will be recognized at the Undergraduate Studies Annual Academic Advising Conference on October 7.

“I received the award notification email on my birthday, so it was a big surprise to me,” Maitland said. “I thank my team in CS Undergraduate Advising, especially Madeline G., for encouraging me and supporting my ideas for the projects.”

According to her nomination materials, Maitland contributed to several initiatives to address student advising needs and improve how the department delivers information and resources.

Among those efforts was a collaboration with UMD’s International Student & Scholar Services to develop additional student support. She also helped design and launch the department’s advising ELMS platform, which brings advising information and resources together in one location.

The nomination noted that Maitland incorporated student feedback into platform updates as the advising team continued developing the resource.

Her work also included changes to the department’s advising newsletter. Maitland developed processes for organizing and distributing the newsletter to provide students with academic, career and advising information in a more consistent format.

For Maitland, interactions with students have played a central role in shaping those efforts.

“It has been great, and I enjoy working with students in the CS program,” Maitland said of her advising experience at UMD. “They are my inspiration to make the advising service better.”

Her selection also reflects work carried out collaboratively within the department’s undergraduate advising office.

“It’s nice to be recognized, but I truly believe the award is for the entire CS Advising team,” she said.

—Story by Samuel Malede Zewdu, CS Communications

CMMRS Marks 10 Years of Connecting Students With CS Research

Samuel Malede Zewdu — Fri, 14 Aug 2026 14:19:12 +0000

The weeklong research school brings together students and faculty from UMD, Cornell and the Max Planck Institutes to explore computer science research and graduate study.

August 14, 2026

The Cornell, Maryland, Max Planck Pre-doctoral Research School in Computer Science (CMMRS) marked its 10th anniversary this summer, continuing a collaboration that introduces undergraduate and master’s students to computer science research, graduate study and research careers.

Held Aug. 3-7 in Saarbrücken, Germany, the weeklong pre-doctoral research school brought students together with faculty and researchers from the University of Maryland, Cornell University and the Max Planck Institutes for technical lectures, mentoring sessions and discussions about careers in academia and industry.

UMD Computer Science Professor Ramani Duraiswami participated in this year’s program, delivering a lecture and joining panel discussions.

CMMRS was established by the three institutions to expose students to current research and give them opportunities to interact with faculty, graduate students and peers from different universities and countries. The program is free for admitted students.

“The basic idea is to keep the best CS undergrads interested in research and, in particular, in our institutions,” Duraiswami said.

Participants attend lectures on current areas of computer and information science and take part in sessions focused on conducting research, applying to graduate school and pursuing careers in academia and industry.

UMD computer science senior David Guan, who participated in this year’s program, said the speakers made their research accessible even to students who entered the sessions without prior knowledge of their fields.

“I was especially surprised by how open and welcoming the faculty were throughout the program,” Guan said. “They often joined us for meals and were always happy to discuss their work or offer advice.”

Guan said another part of the experience was meeting students from different backgrounds and learning about their perspectives.

“Being part of CMMRS was exciting and felt like a rare opportunity to be surrounded by talented students from around the world,” Guan said. “I really enjoyed hearing how their cultures and experiences differed.”

That international exchange was also reflected in the composition of the 2026 cohort, which included students from India, Nigeria, Rwanda, Germany, Bulgaria and the United States.

For participants considering doctoral study, the program also provides opportunities to hear directly from graduate students about their experiences. Connor Clayton, a UMD computer science Ph.D. student and 2026 CMMRS mentor, said the school helps students understand what pursuing a doctorate can involve.

“At CMMRS, students considering a Ph.D. can expect to receive a lot of relevant guidance, including insight about how to decide if a Ph.D. is right for them, what grad student life is like and how to apply,” Clayton said.

—Story by Samuel Malede Zewdu, CS Communications

###

For students interested in learning more about CMMRS, including future program dates, eligibility and application information, visit the CMMRS website.

27 Undergraduates Complete UMD REU Program

Samuel Malede Zewdu — Tue, 11 Aug 2026 16:40:01 +0000

Selected from nearly 400 applicants, the students spent 10 weeks working across seven research teams in computer science and mathematics.

August 11, 2026

27 undergraduate students concluded 10 weeks of research at the University of Maryland’s Department of Computer Science on Aug. 7, completing projects that ranged from artificial intelligence evaluation to mathematical theory as part of the department’s summer Research Experiences for Undergraduates (REU) Combinatorics, Algorithms, and AI for Real Problems (CAAR) program. Selected from nearly 400 applicants, this year’s cohort worked alongside UMD faculty and graduate student mentors across seven research teams.

The REU-CAAR is an NSF-funded research program that brings together students from UMD and universities across the country for an extended introduction to academic research. Participants spend the summer learning the foundations of a research area, developing questions, conducting experiments or theoretical work and presenting their findings.

William Gasarch, a professor of computer science who has helped organize the program since 2013, said students increasingly arrive with prior exposure to research and some familiarity with the subjects they plan to study.

“The cohorts get better and better every year in that students are starting research earlier and earlier,” Gasarch said. “Many of them started doing research in high school, which would have been unusual the first year I ran the program.”

That preparation helped students move from studying existing work early in the summer to investigating questions of their own as the program progressed. The process differed by project, with some teams working on theoretical problems and others examining computational systems and applications.

One group studied Ramsey theory, an area of combinatorics that examines when patterns must appear within sufficiently large mathematical structures. The project led by Gasarch initially focused on understanding established results and developing clearer ways to explain them, including material that could eventually be presented to high school students.

As the group worked through those problems, the project moved beyond its original educational focus and led to new mathematical findings.

Charlie Klawitter, a junior mathematics major at the University of Wisconsin-La Crosse, was part of the Ramsey theory team. He said working alongside students with different backgrounds gave him a broader view of how others approach mathematical research.

“Working with my team was an invaluable experience that provided me with a whole new mindset in the field of research and academia,” Klawitter said. “I learned a lot about the different processes and mindsets of other young academics in the same field as me.”

Klawitter said the experience extended beyond his assigned project. Weekly seminars and organized group activities gave participants opportunities to hear about other research areas and interact with students working on different problems.

Those exchanges were also part of Nghia Truong’s experience. Truong, a junior double majoring in computer science and mathematics, worked on a project led by Professor Jordan Boyd-Graber, examining how artificial intelligence systems respond to challenging questions that combine visual and language information.

The research looks beyond whether an AI system produces a correct or incorrect answer. Instead, the team is studying where systems perform well, where they encounter difficulty and how their capabilities compare with human performance.

Truong said more detailed evaluation methods could give researchers a clearer understanding of what current AI systems can and cannot do.

“More meaningful and fair evaluations can help researchers make more honest claims about AI progress and identify areas that still need improvement,” Truong said. “Ultimately, this work could contribute to building AI systems that are more reliable and work more effectively alongside people.”

Carrying out that research also required students to coordinate their individual responsibilities with the goals of a larger team. Truong’s group met weekly to discuss progress, work through challenges and determine the next steps for the project.

“Working with the group taught me the importance of being organized, communicating clearly and connecting my individual work to a shared research goal,” Truong said.

He also pointed to interactions with students from other institutions as an important part of the summer. Participants entered the program with different academic experiences, giving them opportunities to compare approaches to research and problem-solving.

“Hearing their perspectives and seeing how they approached research broadened the way I think about solving problems,” Truong said.

Across the seven teams, participants learned how research develops from early reading and discussion into more focused investigation. Seminars supplemented that work by introducing participants to faculty research and providing opportunities to discuss graduate study and academic careers.

Gasarch said the program is also intended to expose students to a defining feature of research: unlike coursework, the problem being studied may not have a predetermined solution and can change as new information emerges.

“They will learn how to work in a group and learn that when doing research, the answers are not known, and sometimes the questions change,” Gasarch said.

Other 2026 REU projects:

• Laxman Dhulipala led a project on I/O-efficient parallel algorithms. Students explored theoretical and practical approaches for designing parallel algorithms that reduce the amount of data transferred between main memory and external storage while maintaining high levels of parallelism. The project included developing data structures and algorithms for working with large amounts of data distributed across external memory devices such as solid-state drives.

• Erin Molloy led a computational biology project focused on errors in phylogeny reconstruction. Students examined how errors introduced during genomic data processing can affect algorithms used to reconstruct evolutionary relationships among species or cells. The project included studying error propagation, developing algorithms that are more robust to uncertain inputs and exploring machine learning methods for detecting errors.

• Bill Regli and Bill Gasarch led a project exploring alternative approaches to computing beyond traditional transistor-based systems. Students investigated technologies including memristors, phononic systems, Ising machines and analog computers to understand how they could address computationally difficult problems. The project examined the potential advantages and limitations of these approaches for problems such as Boolean satisfiability, fast Fourier transforms and simulations of physical systems.

• Ming Lin led a machine learning project focused on autonomous driving. Students examined how machine learning systems can use human driving data to better anticipate behavior in environments where autonomous and human-driven vehicles operate together. The project focused on high-risk driving scenarios and applying machine learning techniques to improve understanding of driving dynamics and vehicle safety.

• Sarah Miller led a project exploring alternative neural network architectures for solving games, puzzles and real-world reasoning problems. Students used ideas from combinatorial game theory to build and train neural networks designed to reason efficiently in unfamiliar situations. Depending on their interests, students could extend the work to physical systems, formal verification using Lean or quantum computing while comparing the approaches with transformer-based and reinforcement learning methods.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Algorithms and Theory

TRAILS Awards More Than $515K for Trustworthy AI Research

Samuel Malede Zewdu — Mon, 10 Aug 2026 16:17:58 +0000

The institute, directed by Hal Daumé III, selected five multidisciplinary projects focused on AI literacy, autonomous drones, public trust and local government policy.

August 10, 2026

Helping students learn to evaluate AI. Making autonomous drones safer. Measuring how people express trust in AI. Helping local governments navigate AI policy. Understanding how to reestablish trust after AI failures.

Those are among the questions driving five new research projects selected for the fourth round of seed funding from the Institute for Trustworthy AI in Law & Society (TRAILS), a four-university consortium co-led by the University of Maryland and George Washington University. The institute has awarded just over $515,000 to multidisciplinary teams from UMD, GW, Morgan State University and Cornell University. The grants, ranging from $40,000 to $160,000, are designed to launch new collaborations, generate preliminary results and position investigators to compete for larger external funding.

Supported by a $20 million award from the National Science Foundation and the National Institute of Standards and Technology, TRAILS brings together researchers with expertise in disciplines that include computer science, engineering, law, public policy and the social sciences. Collectively, they are examining one of AI’s biggest challenges: ensuring that AI -infused systems are not only technically advanced, but also trustworthy, accountable and aligned with society’s needs.

The latest round of seed projects spans participatory AI design, methods, sense-making and governance, exploring not only how AI systems work, but also how people use them, how institutions deploy them and how trust is built over time.

“TRAILS is uniquely positioned to connect AI innovation with governance, accountability and public trust at a national scale,” said Hal Daumé III, director of TRAILS and a professor of computer science at UMD. “Over the past few years, we've moved the conversation beyond model accuracy alone to examine how AI systems perform in real-world human, institutional and societal contexts.”

The five projects illustrate how that vision can be put into practice.

“We're continuing to transition from cutting-edge research and scholarship to research with significant impact,” said David Broniatowski, deputy director of TRAILS and a professor of engineering management and systems engineering at GW. “These projects are designed to grow beyond TRAILS through new partnerships and external support.”

Rebuilding Trust After AI Failures

Broniatowski and Cornell’s Valerie Reyna will investigate why people often respond differently to the same AI success or failure. Drawing on reports from the AI Incident Database, the team will examine how people’s values shape their perceptions of AI risks, responsibility and accountability, and whether meaningful explanations tailored to different worldviews are more effective than detailed technical information alone at rebuilding public trust.

Training the Next Public Health Workforce

Morgan State’s Robin Butler and Vishnu Kumar, together with UMD’s Hassanatu Blake, will study how participatory learning influences AI literacy and trust through developing a four-course Certificate in Responsible Public Health AI. Students will help design the curriculum alongside faculty and public health professionals, creating educational resources that emphasize explainability, fairness and accountability while preparing a more diverse workforce to responsibly deploy AI.

Reading Trust Through Language

UMD’s Julia Mendelsohn and Charlotte Vaughn will explore how trust in AI is expressed through everyday conversation. Using computational linguistics to analyze open-ended survey responses, the researchers will identify patterns in how people communicate confidence, uncertainty and concern about AI, creating new tools for measuring public trust across diverse communities.

Safer AI for Autonomous Drones

GW’s Nan Wu and Peng Wei will develop lightweight AI systems that enable autonomous drones to earn users’ trust by making safer predictions in unpredictable urban environments. The project will improve the reliability of vision-language models by reducing decision delays, strengthening performance under challenging conditions and validating the technology through high-fidelity simulations of hazards such as changing light, unexpected obstacles and other real-world conditions.

Helping Local Governments Govern AI

GW’s Joseph Dunne and UMD’s Katie Shilton and Jwan Khisro will examine how local governments translate broad AI principles into everyday practice. Through case studies in Maryland and Washington, D.C., the researchers will explore how public officials define trustworthy AI, incorporate public participation into implementation decisions and navigate accountability challenges, producing practical guidance for cities and counties nationwide adopting AI technologies.

***

In addition to their academic home departments, Hal Daumé III, Julia Mendelsohn and Katie Shilton all have appointments in the University of Maryland Institute for Advanced Computer Studies (UMIACS), which provides administrative and technical support for TRAILS.

—Story by UMIACS

Research Areas:

CS Faculty Join Federal Initiative to Speed AI-Powered Discovery

Samuel Malede Zewdu — Fri, 07 Aug 2026 17:08:19 +0000

August 07, 2026

Four faculty members from the University of Maryland Department of Computer Science, who are affiliated with the University of Maryland Institute for Advanced Computer Studies (UMIACS), are participating in two research projects selected for funding through the U.S. Department of Energy's (DOE) Genesis Mission, a new initiative designed to accelerate AI-driven scientific discovery.

Announced July 22 at the inaugural Genesis Mission Summit in Washington, D.C., the DOE selected 278 projects nationwide to receive a combined $250 million in Phase I funding.

One project includes UMD computer science researchers Chris Metzler, Abhinav Bhatele, and Haizhao Yang, who will help develop an agentic AI framework to overcome computational bottlenecks in the DOE's long-term climate forecasting system. The team aims to improve predictions of floods and droughts, leading to more accurate, localized and long-range forecasts that can help protect critical infrastructure such as power grids and water supplies from extreme weather.

The project, led by Jinwoong Yoo, an assistant research scientist at the University of Maryland and NASA Goddard Space Flight Center, received $750,000 in Phase I funding.

UMD computer scientist Xiaodi Wu is part of a Virginia Tech-led project that will use AI to design more efficient quantum error-correcting codes, helping improve the accuracy and resilience of quantum information. Current approaches require extensive computing resources to protect even a single unit of quantum information, and designing better methods is computationally prohibitive. By using AI to accelerate the design process, the team hopes to develop tools that will speed advances across the quantum computing research community. The project received $720,000 in Phase I funding.

Awardees will gain access to the Genesis Mission Platform, which includes AI agent frameworks, advanced AI models and software from industry partners, and high-performance computing resources at DOE national laboratories and partner facilities. These resources will allow researchers to rapidly design, test and refine new approaches to scientific discovery. Teams selected for Phase II will scale their initial prototypes into tools with broader real-world impact.

—Story adapted from a news brief published by UMIACS

Research Areas:

Learning When to Pivot

Samuel Malede Zewdu — Fri, 07 Aug 2026 14:02:48 +0000

CS alum Ananth Sankaralingam turned early startup experiments into a Y Combinator-backed AI platform for government contractors.

August 07, 2026

Two weeks into an internship at Amazon, alum Ananth Sankaralingam (B.S. ’26, computer science) watched his assigned senior employee leave the company to start a business. The decision stayed with him because it challenged how he had been thinking about work and what he wanted from his own career.

Several years later, Sankaralingam is building Chromie.dev, an artificial intelligence platform for government contractors, from San Francisco. The company was accepted into Y Combinator’s 2026 summer batch after changing direction several times and abandoning an earlier product that failed to attract sustained interest.

For Sankaralingam, the acceptance marked a turning point in a process that began with campus entrepreneurship events, personal software projects and an effort to find work driven by more than compensation.

“He shifted how I viewed work,” Sankaralingam said of his former Amazon mentor. “Instead of going after purely money, he showed me there’s more to it in the sense of following your passion and helping other people.”

Finding Direction

As a student in the University of Maryland Department of Computer Science, Sankaralingam attended events connected to Startup Shell and the Dingman Center for Entrepreneurship. Although he was not accepted into some of the organizations he sought to join, he continued meeting students, researchers and mentors interested in building companies.

Those relationships gave him a community in which he could test ideas and learn from people pursuing similar goals.

“I was increasingly surrounding myself with that entrepreneurship community at Maryland,” Sankaralingam said. “I started forming my own community of friends and mentors in entrepreneurship”

His interest in artificial intelligence developed alongside those early entrepreneurial efforts. A graduate-level course taught by Associate Professor Furong Huang gave him a closer look at large language models and the range of products that could be built around them.

“You realize there’s so much more possible at this moment,” Sankaralingam said. “The opportunity to be at the forefront of that, when the world is changing so drastically, is a once-in-a-lifetime opportunity that I didn’t want to pass up.”

Changing Course

Sankaralingam later co-founded Chromie.dev with Akshay Mistry, a friend from high school. The company initially developed an AI platform that allowed users to create and deploy Chrome extensions.

The product drew little sustained use, and potential customers showed limited willingness to pay. Sankaralingam eventually concluded that something had to change.

“It wasn’t something people really wanted,” he said. “At some point, I got real with myself, and it wasn’t something that I wanted either.”

The decision to move on did not come from a single conversation. Sankaralingam said founders often must evaluate users’ behavior, limited engagement and indirect responses rather than wait for someone to reject an idea outright.

“No one is going to be upfront and direct with you like that,” he said. “You have to reflect on everything you’re doing based on the conversations you’ve had.”

A new direction emerged when Sankaralingam began automating work for a government contractor. He created a system that searched government websites for requests for proposals matching the company’s services.

“I realized this was just the tip of the iceberg,” Sankaralingam said. “There was a larger opportunity underneath in building an AI workspace for government contractors.”

Chromie.dev now helps companies find and assess government opportunities while organizing parts of their business development and capture work.

Shifting Mindsets

Acceptance into Y Combinator brought Sankaralingam and Mistry to San Francisco, but it did not eliminate the daily work of finding customers. Sankaralingam said the program’s main effect was changing how he approached that work.

“You think once you get in, all the customers come knocking on the door,” he said. “That was completely false.”

He moved beyond LinkedIn messages and emails to cold calls, conferences and visits to contracting offices. During one trip, the founders flew to Virginia without booking return tickets and visited about 50 companies, bringing doughnuts and asking employees for brief conversations.

“It didn’t necessarily lead to all the results in the world,” Sankaralingam said. “But it was a mindset shift that I achieved from being around people who were doing whatever needed to be done.”

He now sets aside time each week to review customer conversations, product decisions and advice. The routine helps him assess the company’s direction while measuring progress through the work he can control.

“You can’t guarantee your revenue,” he said, “but you can guarantee your effort and how you’re approaching each day.”

—Story by Samuel Malede Zewdu, CS Communications

Advancing Practical Privacy for Machine Learning

Samuel Malede Zewdu — Wed, 05 Aug 2026 17:35:17 +0000

CS Ph.D. student Yvonne Zhou developed a framework that protects sensitive data during and after model training while reducing training time by up to 15-fold.

August 05, 2026

As artificial intelligence becomes increasingly integrated into sectors such as healthcare, finance and government, organizations face a critical challenge: how to use powerful machine learning tools without compromising confidential data.

At the University of Maryland, fifth-year computer science doctoral student Yvonne Zhou is developing a new framework designed to bridge the gap between privacy-preserving theory and practical, real-world performance. Her work focuses on protecting sensitive information throughout the machine learning pipeline—from the moment data is sent to a cloud server through deployment of the final model.

“Many organizations don’t have the computational resources to train large models themselves and therefore need to outsource training to cloud providers,” Zhou said. “The challenge is how to take advantage of external computing power without exposing private data.”

Zhou’s research addresses what experts consider a major bottleneck in privacy-preserving AI. Existing cryptographic techniques such as fully homomorphic encryption (FHE) allow cloud servers to train models using encrypted data, enabling computation without revealing the underlying information. But the process is notoriously slow and computationally expensive, and it protects data only during training. Once a model is decrypted and deployed, it can still leak sensitive information.

To address this limitation, Zhou developed a framework that combines the rigorous theoretical guarantees of FHE with a customized differential privacy algorithm, which introduces carefully calibrated random noise to help prevent sensitive information from being exposed.

“Together, FHE protects data during outsourced training, while differential privacy protects against information leakage after deployment,” Zhou explained. “In other words, our framework provides end-to-end privacy guarantees.”

A key breakthrough was overcoming the enormous computational cost of combining these two techniques. By developing a new training algorithm that eliminates a particularly expensive computational step, Zhou’s framework achieved up to a 15-fold reduction in training time while maintaining accuracy comparable to significantly slower approaches.

The findings are described in a paper recently presented at the International Conference on Machine Learning (ICML) in South Korea. The work is part of an ongoing collaboration supported by a J.P. Morgan Chase Faculty Research Award and reflects the university’s commitment to working with industry to solve pressing real-world challenges.

“Their research team has a strong interest in privacy-preserving machine learning, so the collaboration brings together academic and industry perspectives on these challenges,” Zhou said. “Although encrypted machine learning is still significantly more expensive than conventional training, improving its efficiency is an important step toward making these technologies practical for applications where trust and confidentiality are essential.”

Zhou's research is co-advised by Dana Dachman-Soled, an associate professor of electrical and computer engineering and a core member of the Maryland Cybersecurity Center, and Min Wu, a professor of electrical and computer engineering. Both hold appointments in the University of Maryland Institute for Advanced Computer Studies (UMIACS).

Dachman-Soled said Zhou quickly mastered entirely new areas while working on the project. Despite having no previous background in fully homomorphic encryption or C++, she was implementing complex encrypted training pipelines within months.

"Beyond her technical abilities, she has consistently demonstrated exceptional initiative and persistence, making her an outstanding researcher and an invaluable member of the team," Dachman-Soled said.

Wu said Zhou's work tackles one of the central challenges in trustworthy AI: balancing strong privacy guarantees with practical performance.

"By combining rigorous theory with algorithmic innovation, she has developed an approach that makes secure machine learning substantially more practical while maintaining strong privacy protections," Wu said.

While the current paper focuses on foundational models that establish rigorous mathematical guarantees, Zhou sees it as an important first step. The team’s next challenge is extending these methods to the deep neural networks that power today’s most advanced AI systems.

“I’m excited to continue exploring how we can make privacy-preserving machine learning both theoretically well-founded and practically useful for the large-scale AI systems that people rely on every day,” Zhou said.

—Story by Diya Sharma, UMIACS communications group

Research Areas:

Graphics Visualization and VR AR

Making Science Easier to See

Samuel Malede Zewdu — Tue, 04 Aug 2026 13:47:52 +0000

Zhicheng Liu and Fumeng Yang received a $559,048 NSF grant to develop tools and design guidelines for more effective science infographics.

August 04, 2026

Scientific discoveries can influence decisions about health, climate and technology, but the research behind them is often difficult to explain outside laboratories and academic journals. Two University of Maryland computer scientists are examining how infographics can present those findings to the general public effectively.

Department of Computer Science Associate Professor Zhicheng Liu and Assistant Professor Fumeng Yang received a $559,048 grant from the National Science Foundation (NSF) to study how science infographics are designed, interpreted and evaluated.

The researchers will study how professional designers choose layouts, images, labels and other visual elements to explain scientific information. They will also examine how those decisions affect what audiences understand, remember and consider credible.

“Our goal is to make it easier for scientists, journalists and designers to communicate scientific discoveries and knowledge through effective and trustworthy infographics,” Liu said.

Infographics are used in news stories, scientific publications and classrooms to explain subjects such as public health, climate change and space exploration. Unlike a conventional chart, an infographic may combine data, diagrams, illustrations, labels and narrative elements in a single visual presentation.

That flexibility can make scientific information more accessible, but it also creates design challenges. Existing software often allows users to arrange text and images without offering guidance on how the elements should work together or whether the finished product communicates its intended message.

“There are few tools or guidelines that help creators design them effectively,” Liu said. “Our project will develop the scientific foundations, software tools and evidence-based design guidelines needed to support the creation of effective science infographics.”

Liu will lead the development of what the researchers call a “grammar” of infographics. Similar to how a language grammar explains how words and sentences fit together, the framework will describe the components of an infographic and the relationships among data, visual elements and storytelling techniques.

The framework will serve as the basis for a JavaScript-based authoring toolkit. The open-source software is intended to help users connect scientific information with visual components and communication goals rather than treating each chart, illustration or annotation as a separate design element.

Yang will focus on how people create and respond to infographics. Her research will examine how designers make decisions, including which information to emphasize, how to organize a visual explanation and what techniques to use to guide a reader through it.

She will also develop a framework for measuring whether an infographic achieves its intended purpose. The evaluation process will consider factors such as what viewers understand, what information they retain and whether they consider the presentation credible.

The project will also produce datasets that other researchers, educators and software developers can use. Yang feels making those resources publicly available could allow others to test the findings, improve the tools and apply them in different areas of science communication.

“Ultimately, we hope to lower the barriers to effective science communication,” Yang said, “helping more people understand complex scientific ideas and make informed decisions.”

—Story by Samuel Malede Zewdu, CS Communications

Research Areas:

Making Complex Math as Easy as Asking a Question

Samuel Malede Zewdu — Wed, 29 Jul 2026 18:58:40 +0000

Haizhao Yang is developing an AI platform that turns everyday questions into mathematical models for complex decision-making.

July 29, 2026

When an emergency like extreme weather strikes, making the right decision in real time can become a matter of life and death. Yet the mathematical optimization tools used to calculate the safest evacuation routes or allocate emergency resources typically require significant computing power and specialized expertise in programming and mathematics.

Haizhao Yang, an associate professor with joint appointments in the University of Maryland's Department of Mathematics and Department of Computer Science, wants to change that. Supported by a $20,000 seed grant from the Artificial Intelligence Interdisciplinary Institute at Maryland (AIM), Yang is developing OptimAI—a cloud-based platform that uses artificial intelligence to make advanced optimization accessible to anyone with a smartphone.

“In real-world applications, not everybody has the expertise to use optimization to find the best solutions, especially for resource allocation,” says Yang, who has an affiliate appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS) and is also a core member of the UMD Center for Machine Learning. “We aim to use AI to lower this barrier so that anyone can make better decisions, particularly in situations related to sustainable decision-making.”

The idea emerged shortly after the release of large language models (LLMs) like ChatGPT. Yang and his team realized that the same natural language capabilities that allow people to converse with AI could also make sophisticated optimization tools available to non-experts.

He points to emergency situations—such as flash floods or rapidly spreading California wildfires—to illustrate the platform's potential. In these moments, people fleeing danger rarely have access to powerful computers. They have only their phones.

With OptimAI, a user could simply describe their surroundings by speaking into their device. The AI would interpret that natural language, incorporate real-time environmental data from sources such as radar and satellites, and automatically generate a mathematical optimization model to calculate the safest, most efficient evacuation route—all without requiring users to install specialized software or understand the underlying mathematics.

Building a system capable of generating reliable, life-saving recommendations requires combining rigorous mathematical theory with cutting-edge computer science, an intersection where Yang has extensive expertise.

Because LLMs are prone to hallucinations—producing confident but incorrect responses—using them in safety-critical situations presents significant challenges. Yang's team plans to ground the AI's outputs in mathematical theory, using it as a source of truth to improve the system's reliability.

“This is interdisciplinary research that really relies on math to make the AI tool trustworthy,” Yang explains. “With mathematical theory, we may be able to develop very robust, trustworthy AI to reduce hallucination.”

At the same time, the platform's computer science framework must seamlessly translate spoken language into executable optimization code, run those computations on cloud-based hardware and return results to users within seconds.

The one-year AIM seed grant will primarily support student researchers, including undergraduate computer science students affiliated with UMIACS, who will help test and refine the platform's core functionality. The team will also leverage UMIACS's GPU clusters to train and run computational models.

Yang hopes the project's early results will position the team to pursue larger grants from the National Science Foundation. Ultimately, he envisions a future in which mathematical optimization is no longer limited to researchers and specialized industries but is available to anyone who needs to make better decisions—simply by asking a question.

—Story by Diya Sharma, UMIACS communications group

Research Areas:

Bioinformatics and Computational Biology

CS Faculty Take Leading Roles at Global Conferences

Samuel Malede Zewdu — Mon, 27 Jul 2026 14:57:06 +0000

Mihai Pop and Philip Resnik delivered keynote addresses and held leadership roles at international conferences focused on artificial intelligence and computational biology.

July 27, 2026

University of Maryland computer science faculty members Mihai Pop and Philip Resnik joined fellow UMIACS faculty member Joshua Weitz in taking on leadership and speaking roles at major international conferences in July 2026. Their participation highlighted research across artificial intelligence, computational biology and microbiome science.

Philip Resnik, a professor of linguistics and an affiliate professor of computer science, delivered the opening keynote address at the Annual Meeting of the Association for Computational Linguistics (ACL 2026) in San Diego. His presentation, A New Balancing Act: Reflections on the Relationship Between Computational Linguistics and AI, argued that the commercial success of AI and large language models (LLMs) has disrupted the balance of priorities in the computational linguistics community. He outlined three key challenges facing the field and encouraged researchers to approach them through constructive dialogue rather than opposition.

“I wanted to bring conversations already happening across the field into the open and frame them in a way that encourages productive discussion,” Resnik said. “This community played a pivotal role in creating the AI revolution and I argued that we’ve got a crucial and distinctive role to play in how AI evolves moving forward.”

Mihai Pop, an MPower Professor of Computer Science, served as co-chair of the 2026 Conference on Intelligent Systems for Molecular Biology (ISMB) in Washington, D.C. He also delivered the Distinguished Keynote Address at the ISCB Youth Bioinformatics Symposium, a companion event that introduces middle and high school students to computational biology and bioinformatics.

In his keynote, My Long and Winding Road to Computational Microbiome Analytics, Pop reflected on the unexpected path that led him to computational microbiome research, sharing lessons from both successes and setbacks before offering advice to students beginning careers in science.

Joshua Weitz, a professor of biology and the Clark Leadership Chair in Data Analytics, co-chaired the Bioinformatics and AI sessions at the 2026 Viruses of Microbes Conference in Prague, Czech Republic. The conference explores the impacts of viruses that infect microbes, from molecular mechanisms and evolution to ecology and therapeutic applications. Weitz opened the sessions with a presentation titled Computational Models of the Eco-Evolutionary Dynamics of Inefficient Viral Infections.

“The session highlighted how distinct modeling approaches—from mechanistic simulations to AI-driven inference—can be used to understand how viral infections of microbes impact human and environmental health,” Weitz said. “International in scope, the session reinforced the centrality of collaboration to move this field forward.”

—Story by Maria Herd, UMIACS communications group

Research Areas:

UMD Computer Science Graduates Report 95% Job Placement Rate

Samuel Malede Zewdu — Wed, 22 Jul 2026 14:48:09 +0000

The University Career Center survey shows strong employment outcomes, higher starting salaries and continued graduate school enrollment for the class of 2025.

July 22, 2026

The University of Maryland’s Department of Computer Science reported a 95% job placement rate for its class of 2025, according to a survey conducted by the University Career Center. The annual graduation survey, which covers students who completed degrees in August and December 2024 and May 2025, found that graduates continued to secure employment across the technology sector while also pursuing advanced degrees at universities nationwide.

"These results reflect both the strong technical foundation our students develop and the many ways they engage with research, internships and other opportunities while they're here," said Department Chair Matthias Zwicker, who holds the Elizabeth Iribe Chair for Innovation and the Phillip H. and Catherine C. Horvitz Professorship. “It's rewarding to see the strong career outcomes our graduates continue to achieve.”

The survey tracked outcomes for 1,188 graduates and collected data from 906 students, resulting in a 76% knowledge rate. Among respondents, 78.5% entered the workforce and 16.2% continued their education. Only 1.5% of graduates were reported as unplaced at the time of the survey, while 3.8% had not yet finalized their post-graduation plans.

Graduates also reported higher starting salaries than the previous year. The median starting salary increased to $119,000, with the middle half of respondents earning between $90,000 and $145,000. More than 200 graduates also reported receiving employment bonuses, with a median bonus of just over $10,300.

Internships continued to be one of the strongest predictors of employment. More than 85% of respondents completed at least one internship during their time at Maryland, and nearly 60% completed multiple internships. Among graduates who described how they found their jobs, internships were the most frequently cited pathway, ahead of online job sites, personal contacts and career fairs.

Graduates accepted positions with employers across the technology, finance, government and defense sectors. Companies hiring multiple UMD graduates included Amazon, Capital One, Meta, Bloomberg, Google and Microsoft, while others joined startups, federal agencies and research organizations. Software engineering remained the most common career path, with graduates also entering roles in artificial intelligence, cybersecurity and data science.

Although many graduates accepted jobs across the country, nearly 45% remained in Maryland, Virginia or Washington, D.C. Others relocated to technology hubs including California, New York and Washington state.

Graduate school also remained a popular option. Of the 146 graduates who continued their education, nearly nine in 10 enrolled in master's programs, most commonly in computer science. More than half chose to remain at UMD for graduate study.

“Our students define success in many different ways,” Zwicker said. “Whether they enter the workforce, continue their education or pursue entrepreneurial ventures, we're pleased to see them building on the experiences they had at UMD.”

—Story by Samuel Malede Zewdu, CS Communications

Three CS Terps Received CMNS Alumni Network Endowed Undergraduate Awards

Samuel Malede Zewdu — Tue, 21 Jul 2026 18:11:32 +0000

Raghav Thind, Nghia Truong and Priya Tyagi received funding to support summer research, conference travel and career development.

July 21, 2026

Every year, the College of Computer, Mathematical, and Natural Sciences (CMNS) Alumni Network offers awards to help undergraduates defray the costs of conducting research, attending conferences or interning over the summer.

Read below how this year’s award recipients are furthering their career development with funding from the CMNS Alumni Network Endowed Undergraduate Awards program.

Raghav Thind

Since becoming a senior computer science major at UMD, Raghav Thind has been drawn to research and development. He is using this award funding to attend the prestigious Summer@EPFL program, a three-month summer research fellowship hosted by the Swiss Federal Technology Institute of Lausanne in Switzerland.

“Not only will this award impact my outcome this summer, but it will also play an important role in helping me to continue developing as a research scientist and make me capable of contributing to impactful, real-world innovation,” Thind said.

After graduation next year, Thind plans to pursue a Ph.D. to become a research scientist “at the forefront of technological innovation.”

Nghia Truong

Junior computer science and mathematics double-degree student Nghia Truong used his award funding to travel to present his first-author paper at the 2026 IEEE International Conference on Acoustics, Speech, and Signal Processing in Barcelona.

This summer, Truong will continue his research with UMD Mathematics Associate Professor Yun Yang on the theoretical foundations of transformer attention and its connections to kernel methods. He will also participate in a Research Experience for Undergraduates (REU) with Computer Science Professor Jordan Boyd-Graber focused on building a multimedia quiz system that challenges artificial intelligence systems.

“As an international student conducting research without a stipend, this scholarship has made a significant difference by helping alleviate financial burdens,” Truong said. “Beyond that, it is deeply encouraging and motivates me to continue pursuing research with dedication and purpose.”

After graduation, Truong hopes to pursue a Ph.D. in mathematics and machine learning theory and looks forward to making meaningful scientific contributions and mentoring the next generation of students.

Priya Tyagi

Receiving a total hip replacement at age 17 due to slipped capital femoral epiphysis inspired senior computer science major Priya Tyagi to pursue computational solutions to biological problems.

Tyagi is using her award funding to continue her research in Bioengineering Assistant Professor Alexander Xu’s Spatial Biology Lab. Her project examines whether atomic force microscopy tissue measurements could be replaced by computational predictions trained on imaging data—thereby saving both time and money. Tyagi is focusing on the pipeline’s image processing, feature engineering and machine learning components to accelerate this research.

After graduation, Tyagi plans to pursue an M.D.-Ph.D. in computational biology.

“The goal is to become a physician-scientist,” Tyagi said. “I am interested in using computational approaches to uncover medical mysteries about diseases that clinical practice alone cannot address, and eventually translating those discoveries into clinical interventions.”

—Story by CMNS

Research Areas:

Michelle Mazurek Builds People-Centered Security Tools

Samuel Malede Zewdu — Mon, 20 Jul 2026 17:56:02 +0000

She discusses her path into computing, her work on privacy and data de-identification, and why communication matters in human-centered security research.

July 20, 2026

University of Maryland Professor of Computer Science Michelle Mazurek studies how people navigate security and privacy decisions in everyday digital systems. Affiliated with the University of Maryland Institute for Advanced Computer Studies (UMIACS) and the Maryland Cybersecurity Center (MC2), her research connects technical questions with the human choices behind them, from how software developers handle secure coding to how researchers protect sensitive data before sharing it.

In this Q&A, Mazurek discusses the early experiences that drew her to computing, her work on data de-identification and why communication is central to human-centered security research.

Was there a defining moment that shaped your path into computer science?

There was more than one. Growing up, my dad was an electrical engineer, and he taught my sister and me that engineering was a way to solve problems, whether something broke in the house or needed to be fixed on a computer.

I first tried programming in second or third grade with Logo, where you could move a pointer on a screen by writing commands. Later, we used Lego Logo, an early version of Lego robotics systems. We built things with Legos and programmed motors to make them move. That was probably how I became interested in programming and computer science.

Can you describe your research focus?

My work is at the intersection of security and privacy with human-computer interaction. We study how people make decisions related to privacy and security and how to help those decisions go better.

That includes end users trying to protect themselves, software developers whose code affects others and professionals such as librarians who manage privacy issues for the people they serve.

What is one current project you are working on?

One project I am working on with my student Wentao Guo focuses on disclosure avoidance and de-identification for research data.

Social science and health researchers often collect human subjects data about topics such as health outcomes, poverty, income or personal background. There is interest in releasing that data after a study so other researchers can replicate findings or conduct new analyses.

At the same time, there are risks. If data can be re-identified, someone could learn sensitive information about the people who participated. Researchers have a responsibility to protect participants, especially when they have promised that study data will not be tied back to a person’s identity.

The challenge is that many things besides a name can identify someone. A combination of details, such as job title, gender, department or years at an institution, can quickly narrow the group of possible people.

How is your group approaching that problem?

There are technical approaches to making data less identifiable, but they involve trade-offs. If you replace all the data with random numbers, it may be private, but it will not be useful for research.

Many social science and medical researchers are expected to de-identify and release data, but they often do not receive detailed training on how to do it. We have talked to researchers and reviewed existing documentation to understand their goals and constraints.

More recently, we studied how people use tools designed to support de-identification. We are interested in tools that help researchers understand the choices they are making and select an approach that fits their data and research goals.

Can you describe your lab?

My group works closely with the Maryland Cybersecurity Center (MC2), where I am affiliated. At the moment, the lab includes a postdoc, five doctoral students, and a couple of undergraduate students.

That size allows us to work on several projects while still giving each student the attention they need.

What role does MC2 play on campus?

The Maryland Cybersecurity Center brings together faculty and students working on security and privacy research across campus. It includes about 10 to 15 faculty members, mostly from computer science and electrical and computer engineering, with some participation from other departments.

The center gives researchers a way to communicate, share resources and connect students with a broader community working on related problems.

How does your work connect to broader issues in computer science?

Security is a technical problem, but it is also a human and organizational problem. Secure software matters, and the challenge is growing as software systems become larger and more complex.

In many cases, the issue is not that no solution exists. It may be that a vulnerability was not noticed soon enough, a patch was not applied in time or the process did not work at scale. Technical solutions are important, but we also need to understand the barriers that prevent those solutions from being used effectively.

What inspired you to join the University of Maryland?

I was an undergraduate at Maryland, where I studied electrical and computer engineering. My husband is also an alum, and many of our friends and family connections are in the area, so returning to Maryland felt like coming back home in some ways.

It was also a chance to see the university from a different perspective, as a professor rather than a student. The computer science department spans many areas, including human-computer interaction and security.

What advice would you give students interested in your research area?

Students should think carefully about communication skills. Technical skills and programming are important, but this kind of work also requires understanding people, asking questions and explaining findings clearly.

You need to be able to talk to people about the challenges they face, write about what you find and communicate ideas to people who may build new systems. Those skills are important in human-centered security and privacy research.

—Story by Samuel Malede Zewdu, CS Communications

Research Areas: