Summary
Guilhem Piat is a postdoctoral researcher based in Paris with 11 years of experience focused on advancing language-model capabilities beyond scale-driven gains. Since his PhD began in 2019 he has explored ways to let Transformers leverage dense, structured knowledge—essentially teaching models to "look things up" rather than infer everything from context alone. He has combined research and teaching (functional programming, OOP, and AI) at leading French universities and now continues this work at IMT Atlantique. Guilhem’s perspective is practical and inquisitive: he sees recent tools like Copilot and ChatGPT as milestones toward AGI but is driven to find more efficient, principled architectures. He also shares compact code experiments on GitHub, reflecting a hands-on approach to prototyping ideas alongside formal research.
11 years of coding experience
Master of Science - MS, Machine Learning & Data Science, Master of Science - MS, Machine Learning & Data Science at Univ Paris Diderot
Université Paris-Saclay