AI Scientist
Contact: first.last@mila.quebec
Full Bio - CV - Publications - Collaboration
Kellin Pelrine is a research scientist at FAR.AI and a PhD candidate at McGill University and the Mila AI Institute (supervised by Reihaneh Rabbany and Jean-François Godbout). Previously he cofounded Stitch, Mutual, and Machine Learning Sports, and before that he completed a master’s in economics at Yale University and a master’s in applied mathematics at the University of Colorado at Boulder.
Pelrine leads cross-functional teams of over 40 collaborators hoping to create practical, transformatively impactful machine learning systems. His work focuses on making generative AI a reliable, trustworthy tool to navigate the information ecosystem, and on exposing, understanding, and solving misuse risks of frontier models. Pelrine has worked on diverse topics spanning AI security, education, persuasion and manipulation, political polarization, human trafficking, adversarial robustness, and temporal graphs. Projects he has directed while a grad student have received multi-million dollar funding. His research has been published in top venues such as NeurIPS, ICML, AAAI, EMNLP, and WebConf, featured multiple times in the news such as the Financial Times, The Times, Vice, Scientific American, and Nature Magazine, and cited in the US Senate.
When schedule allows, he is a foodie, and an avid player and teacher of the board game Go. He has twice represented the US in international competition, including the World Mind Sports Games. For beating “superhuman” Go AIs as a human, he has been called the man who beat the machine.