EECS students win IBM Granite Pioneer Award in inaugural AI racing competition
A team of five students from the School of Electronic Engineering and Computer Science at Queen Mary University of London has been named one of just two winners of the IBM Granite Pioneer Award, following their participation in the inaugural IBM AI Racing League.
The team, known as The MonDragons, comprised Kishal Chhetri, Arnav Dubey, Khaled Alompeint, Syed Mustafa Saeed Ainuddin, and Tamzeed Araf Islam. Over the course of the competition, which ran from December 2025 to February 2026, the team developed an autonomous AI driver capable of navigating the Corkscrew racetrack in TORCS - an open-source racing simulator - achieving a fastest lap time of 1:47.84 seconds.
The IBM AI Racing League challenged participants to build and optimise an AI that could drive an F1 car around a virtual racetrack as quickly as possible. As one of the very first cohorts to take part in this new competition, the teams' feedback will also help shape how the competition evolves in future editions.
The MonDragons' AI is based on a rule-based approach - similar in principle to the behaviour of AI characters in video games. Rather than using complex machine learning models, the team defined a set of driving rules, such as slowing down when approaching a corner, and then refined those rules iteratively based on real-time sensor data fed back from the TORCS game engine. The output of their Python script was transmitted directly back to the simulator, allowing the virtual car to react and adjust its behaviour accordingly. This process of devising a new strategy, implementing it and testing it rigorously was central to the team's development journey.
The MonDragons were awarded the IBM Granite Pioneer Award for demonstrating exceptional technical ability, embodying the spirit of community, and embracing the creative nature of the competition, alongside generous prizes.
The team would like to thank John McNamara for his support in organising the competition, and Professor Steve Uhlig for his continued encouragement throughout their involvement.
The team has documented their full development journey, including the challenges and setbacks they encountered along the way, in a blog post available here.
The next phase of the IBM AI Racing League is now open, with registration available until 31 May 2026. Students interested in taking part can find out more and sign up here.