Summary
Jacques Thibodeau is a founder and machine learning engineer with nine years of experience focused on reducing risks from advanced AI through hands-on experiments, model editing, and automated evaluation pipelines. Based in London and trained in physics, he blends deep technical skills in deep learning, NLP, and computer vision (PyTorch, Huggingface) with practical cloud experience on AWS and Azure. He has led startups and research teams—co-founding Coordinal Research and now running a stealth venture—while contributing novel mechanistic interpretability insights such as identifying the "Reversal Curse" prior to wider publication. Jacques builds tooling to automatically discover new misbehaviors in model updates and emphasizes reproducible, unsupervised evaluation methods that support safer R&D and red-teaming. An educator and former government data scientist, he brings a systems-thinking approach to alignment work and frequently shares ideas publicly on Twitter and GitHub.
10 years of coding experience
8 years of employment as a software developer
Master's degree Physics, Master's degree Physics at Université de Moncton
Full Stack Deep Learning
Machine Learning Engineer Nanodegree with Amazon Web Services (AWS) Machine Learning, Machine Learning Engineer Nanodegree with Amazon Web Services (AWS) Machine Learning at Udacity
Artificial Intelligence, Artificial Intelligence at AI Safety Camp (Building tool with GPT-3)
French, English