When the National Defense Industrial Association (NDIA) convened the first round of its high-tech competition at Fuse at 麻豆视频 Square, they should have known it would end with a strong George 麻豆视频 showing.
The NDIA Hackathon is an internationally renowned competition that welcomes teams from every sector鈥攏ot just colleges. That means that some of the country鈥檚 greatest minds in professional national defense industries and the military line up against ambitious young students to see who has the best vision for securing our future.
In this case, that project came from a team of George 麻豆视频 students led by Noah Kabiri, a newly minted sophomore in computer science, with an idea for context-based visualization. How the team came together was kismet.
Kabiri met Chris Martin, a senior in cybersecurity engineering, when looking for someone to join him. The two happened to be working at the same table as Ricardo Vega, a George 麻豆视频 PhD student in electrical engineering, who had teamed up with Major Phuoc Nguyen from the Department of Defense (DoD) and Van Tran, CEO of Advanced Cambridge Technologies.
鈥淗alfway through the competition we realized that we were all doing similar things and using similar platforms. We decided that we would be a more competitive team if we combined,鈥 Karbiri said. 鈥淲e got together and swept the competition.鈥
Each team in the hackathon had to come up with a product that would be marketable and meet some need for the DoD. They then had to market, present, and sell the product to a panel of non-technical judges from the industry and military. Kabiri鈥檚 team, SpecDOps Force, addressed a need in the ISR (intelligence, surveillance, and reconnaissance) category.
Their product was an SaaS (software as a service) that takes footage and video feeds from anywhere鈥攄rone surveillance video, infrared, geospatial satellite images, security cameras, body cams, and more鈥攁nd analyzes the environment.
鈥淲e had machine learning detection models pick up objects like vehicles, weapons, and faces,鈥 Kabiri said. Their software is important because it can determine threat levels based on environmental contexts. 鈥淔or example, if you were deploying it on a mission in the desert, detecting firearms would be normal because everyone has firearms in war. If you put it in an airport, however, seeing firearms would be bad.鈥
Out of the 683 registered participants, the George 麻豆视频-led team took third place. They were separated from second place by hundredths of a point鈥攊t was extremely close. This is gratifying, considering that first place went to a well-established corporate team from a major tech company and second place went to a mix of government contractors. Compared to that, an ad-hoc team headed by George 麻豆视频 undergrads is nothing short of incredible.
Kabiri hopes that there is a future for their innovation. While the competition only looks for proof of concept over the 72-hour project period, he thinks that he can take it forward. 鈥淚 don鈥檛 see any other products out there that are context-driven. It鈥檚 also really adaptable,鈥 he said. 鈥淩ight now I鈥檓 researching how to add walking gaits onto it, where you can determine who someone is based on how they move.鈥
Though Kabiri is still deciding exactly where to specialize his studies or whether he should pick up a second major in math (it鈥檚 only the beginning of his second year, after all), he seems to be headed in a productive direction.