Maxime Chevalier-boisvert
Principal Compiler Architect at yasp
Montreal, Quebec, Canada
Summary
🤩
Rockstar🎓
Top SchoolMaxime Chevalier-Boisvert is a Principal Compiler Architect with a PhD in compiler design and 15 years of experience building high-performance systems at the intersection of compilers, AI, and 3D graphics. He led the team that created the official Ruby JIT (YJIT) and architected subsequent next-generation JIT work adopted across major platforms, demonstrating impact at companies like Shopify and clients such as GitHub, Stripe, and Discourse. His research-driven engineering at Mila produced code cited in over 1,200 papers and informed practical RL and sim-to-real tooling used in open-source projects like MiniGrid and BabyAI. Comfortable from LLVM-based GPU compilers to reinforcement-learning environments, he combines deep technical authorship with effective technical leadership and clear communication. Based in Montreal, he enjoys blending creative problem solving with hands-on contributions across graphics, ML, electronics and music—skills that often surface in unexpected cross-disciplinary solutions.
15 years of coding experience
4 years of employment as a software developer
Ph.D. Computer Science, Ph.D. Computer Science at Université de MontréalMaster's Degree Computer Science, Master's Degree Computer Science at McGill University
English, French
Github Skills (30)
environmental10
caching10
pytorch10
dev-environment10
data-pipelines10
python10
gymnasium10
openai-gym10
machine-learning10
reinforcement-learning10
enviroment10
pyglet10
development-environment10
deep-reinforcement-learning10
data-pipeline10
Programming languages (13)
Github contributions (5)
Farama-Foundation/Minigrid
Dec 2017 - Mar 2022
Simple and easily configurable grid world environments for reinforcement learning
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:298 commits, 65 PRs, 322 pushes in 4 years 3 months
Contributions summary:Maxime contributed to the MiniGrid environment by implementing features, fixing bugs, and adding new environments. Their work included fixing issues with the basicrl code and addressing bugs related to the exitDoorPos. They also added features such as keyboard handling, object position tracking and the ability to generate colored walls and floor tiles. Additionally, the user reorganized files and added documentation.
grid-worldconfigurablereinforcement-learningdeep-reinforcement-learningreinforcement
mila-iqia/babyai
Jul 2018 - Mar 2021
BabyAI platform. A testbed for training agents to understand and execute language commands.
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:3 reviews, 263 commits, 82 PRs in 2 years 8 months
Contributions summary:Maxime primarily contributed to the BabyAI platform, a testbed for training agents to understand and execute language commands. Their contributions included adding new features and debugging existing levels. The user demonstrated proficiency in language processing by implementing different instructions and verifying agent behaviors. The added new debug versions of the game levels, including enhancements to the OpenDoorsOrder level.
nlpagentsmachine-learningimitation-learningnlp-machine-learning
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.