Waris Radji is a founding engineer and PhD student in Theoretical & Deep Reinforcement Learning based in Paris, combining eight years of hands-on software and research experience. He specializes in Scala and machine learning, contributing to high-profile open-source projects like scalameta/metals and PyTorch's rl library where he improved language-server features and reinforcement-learning tensor specs. His background spans research roles at Inria and practical engineering at labs such as CNRS and LIPN, where he implemented bio-inspired clustering algorithms and built ML notebooks with Spark. Equally comfortable in academic and startup settings, he co-founded his university AI club and now helps build a startup while pursuing doctoral research. An unusual blend of strengths—deep theoretical RL work alongside developer tooling and production-level Scala—drives his goal to complete a PhD in machine learning. He’s also a competitive Street Lifting athlete, reflecting discipline and a taste for structured, progressive training.
8 years of coding experience
3 years of employment as a software developer
DUT Informatique, Computer Science, DUT Informatique, Computer Science at Université Sorbonne Paris Nord
Engineer's degree, Computer Science, Engineer's degree, Computer Science at ENSEIRB-MATMECA
Numerics Graduate Program, Optimisation and Decision, Numerics Graduate Program, Optimisation and Decision at Université de Bordeaux
Doctor of Philosophy - PhD, Mathematics and Computer Science, Doctor of Philosophy - PhD, Mathematics and Computer Science at Université de Lille
Contributions:18 reviews, 30 commits, 5 PRs in 4 months
Contributions summary:Waris primarily contributed to the `scalameta/metals` project by implementing features related to the language server. They added file name completions for Scala 3 and refactored code to utilize `toLsp` instead of `toLSP`. The user also made changes to the call hierarchy, adding function signature details and fixing issues. Their work focused on enhancing the functionality of the language server, specifically improving code completion and call hierarchy features.
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.
Role in this project:
ML Engineer
Contributions:17 reviews, 6 commits, 7 PRs in 10 days
Contributions summary:Waris primarily contributed to the PyTorch/RL repository by implementing and refining components related to reinforcement learning environments and tensor specifications. Their work involved removing obsolete classes, adding a step counter transform for environment control, and introducing a `MultiDiscreteTensorSpec` with enhanced functionality. They also made bug fixes and code quality improvements, particularly to tensor specifications and transforms, including the `MultOneHotDiscreteTensorSpec`.
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.