UMD CS Majors Develop AI System to Improve School Safety

Concerns over school safety have grown in recent years as threats and incidents of gun violence continue to affect educational communities nationwide. While schools have adopted drills and surveillance systems to improve preparedness, many still seek more effective ways to prevent and respond to potential dangers. At the University of Maryland, a group of students is exploring how artificial intelligence can help close that gap.

Their work culminated in a first-place finish at the 2025 Xperience Competition for their project, DefenX, an AI-based gun detection and alert system designed to enhance school safety. The competition awarded them a $250,000 investment to help develop and deploy the system.

DefenX detects, tracks and alerts relevant stakeholders—including law enforcement—when an active shooter threat is detected. Beyond detection, it provides tools for post-incident mental health support to students and staff. The team developed the system over a year and a half, integrating it with existing school infrastructure to make adoption easier and more cost-effective.

The winning team from the Department of Computer Science includes senior computer science majors Arnav Dadarya, Srinidhi Gubba, Avishi Gupta, Smithi Mahendran and Niyant Patel; applied machine learning master’s student Nithin Skantha Murugan; and Arundhati Ghose (M.S. ’25, applied machine learning).

For Gupta, the victory was both emotional and validating after months of intensive work.

“It just felt like fireworks were bursting behind me,” she said. “This has been part of my life since freshman year. To see all our efforts finally work out felt amazing, like we can finally make the difference we were talking about. We have the backing now to start pilot testing with schools and actually make an impact in students’ lives.”

Building from Scratch

The team’s journey began in the fall of 2023, when they attended an information session about the Xperience Competition hosted by UMD’s xFoundry. The challenge of designing a technology to improve public safety struck a chord with the group, particularly because of their shared experiences as students.

“School safety is something that affects all of us,” Gupta said. “My high school had safety threats, and even though nothing was carried out, it stayed with me. Many of my early projects were centered on school safety and gun violence, so this just felt fitting.”

That personal connection helped drive the team to pursue a solution rooted in both empathy and practicality. The students also saw the competition as an opportunity to apply technical skills to a real-world problem with urgent social relevance.

DefenX

DefenX operates as an end-to-end system for schools. Using AI-driven image recognition, it detects firearms in real time through existing surveillance cameras. Once a potential threat is identified, the system tracks the individual’s movement and sends alerts to students, teachers, administrators and, when appropriate, law enforcement.

The technology integrates with schools’ current infrastructure, reducing the need for new hardware installations. Teachers can receive notifications through wearable devices designed to alert them even when they’re not near their phones, allowing them to initiate lockdown procedures immediately.

Unlike other security tools, DefenX extends beyond incident detection. The system includes features for post-incident recovery, such as mental health resources and an interactive chatbot that allows schools to review and understand incident data for future preparedness.

“One of our mentors, a researcher, encouraged us to think about what happens after an incident,” Gupta said. “That led to the post-incident support component, which we saw as just as important as detection.”

AI as the Core

Artificial intelligence lies at the heart of the system’s design. The team relied on AI for its speed and precision in detecting threats, capabilities that can be critical during emergencies.

“AI is essential because of how quickly it can analyze,” Gupta said. “Most school shootings last only a few minutes. The faster you can detect and alert, the more lives you can protect. AI helps us achieve that speed.”

By automating detection and response, the system reduces reliance on human reporting, which can be slower or inconsistent during crises.

Interdisciplinary Collaboration

A diverse mix of academic backgrounds strengthened DefenX’s design and development. While most of the team were computer science students, it also included members with pre-law and finance experience. 

According to Dadarya, this interdisciplinary approach helped shape both the technical and strategic aspects of the project.

“Having a multidisciplinary team was instrumental to our success,” he said. “Team members with diverse backgrounds enabled us to design a product that not only met legal compliance requirements but also remained financially accessible to schools.”

By dividing responsibilities based on expertise—ranging from backend engineering to legal compliance and financial modeling—the team was able to work in parallel and maintain momentum under tight deadlines. Regular meetings ensured that technical, ethical and business considerations remained.

From Campus to Community

With the $250,000 investment secured, the team is now preparing to pilot DefenX in local schools, beginning with potential partners in Prince George’s County. 

“We’ve tested the system in the IDEA Factory,” Gupta said. “Our next step is real-world pilot testing in schools. We’re meeting with administrators, attending networking events and looking for opportunities to partner with districts interested in testing the technology.”

While the road ahead includes regulatory, ethical and logistical hurdles, the team sees the experience as a chance to contribute meaningfully to an ongoing national conversation about school safety and the role of technology in prevention.

“We’re students building for students,” Gupta said. “We’ve lived through the fear of safety threats, and we know how important it is to feel secure at school. Our goal now is to keep improving DefenX through pilot testing and iterations until it’s ready to make that difference.”

—Story by Samuel Malede Zewdu, CS Communications