A pair of computer science and neuroscience students at Northeastern University gave the phrase “playing like a robot” new meaning.
Second-year students Nikhil Mukraj and Jacob Shechter recently unveiled a robot that plays live poker at the school’s RISE 2025 Expo. The gathering featured hundreds of student researchers across several disciplines, showing off their work.
According to the report from Northeastern Global News, Mukraj’s and Schecter’s robot stole the show.
Unlike some previous computerized poker-playing efforts and online poker bots, the students’ robot actually functions like a real-life player and can participate in live poker games. The goal was to improve “pick-and-place” robotics algorithms for industrial uses.
“In various manufacturing and warehouse settings, you want to have precise control over a given object or item,” Mukraj said. “Robot’s motions need to be very precise.”
In this case, that meant having the robot control playing cards and chips while implementing some poker strategy. The game seemed like a perfect mix of precision movements and human logic, offering the challenge of designing a robot that could play the physical game while also making solid strategic plays at the table.
Mukraj and Shechter used a modular design with parts made from a 3D printer. The robot handles the cards via an arm with a suction cup and has a full range of movement. A card recognition algorithm was included to allow the robot actually to play and participate in a game.
“Poker is a semi-random game — you don’t always know what’s going to happen next,” Mukraj said. “The algorithms that we developed to solve poker can be applied to other unpredictable environments. We see this coming up quite a bit in navigation.”
Some of those possible uses for the technology could include manufacturing to self-driving cars, the project builders said.
Mukraj and Shechter hope to enhance their bot even more so that it can also deal cards.
Several high-stakes pros have previously squared off against poker-playing computers, including Phil Laak and Ali Eslami beating Polaris in 2007, which was designed by the University of Alberta’s Computer Poker Research Group. The group later improved on that with Cepheus, using artificial intelligence for an even more advanced program.
However, those computers were built to try to play nearly unbeatable poker. In this case, the two Boston-area students were not focused on strategy. Instead, they zoned in on the machine’s robotics.
