Nathan Lichtlé
Graduate Student
Berkeley, California, United States
Summary
👤
Senior🎓
Top SchoolNathan Lichtlé is a PhD researcher in EECS at UC Berkeley and a co-founder and Chief Scientist at Yumi Health, bringing eight years of applied AI and back-end engineering experience to healthcare and traffic/decision-making systems. He blends deep academic training in applied mathematics and AI from top French and US institutions with practical contributions to high-profile open-source projects like DeepMind's OpenSpiel and the Flow traffic RL framework, where he fixed subtle game-state and API integration bugs. As an AI Research Scientist at Layr, he applies rigorous RL and systems thinking to product-ready research, bridging simulation fidelity and production constraints. Known for cleaning up “magic numbers,” improving action/state representations, and smoothing complex API integrations, he excels at turning theoretical models into reliable engineering. Based in Berkeley, he mixes entrepreneurial drive with reproducible research practices, making him effective at both lab-scale innovation and shipping robust backend systems.
8 years of coding experience
Bachelor and Master's degree, Computer Science Department, Bachelor and Master's degree, Computer Science Department at École normale supérieure Paris-SaclayFrench preparatory classes (CPGE), MPSI and MP, Mathematics and Computer Science, French preparatory classes (CPGE), MPSI and MP, Mathematics and Computer Science at Lycée Jean-Baptiste Kléber
Github Skills (18)
c-language10
python10
apidoc10
coregames10
reinforcement-learning10
videogames10
api10
cprogramming-language10
congestion-control9
multi-agent9
controls9
traffic9
algorithm7
algorithms7
git7
Programming languages (10)
Github contributions (5)
flow-project/flow
Mar 2019 - Apr 2021
Computational framework for reinforcement learning in traffic control
Role in this project:
Back-end Developer
Back-end DeveloperContributions:597 commits, 47 PRs, 265 pushes in 2 years 1 month
Contributions summary:Nathan primarily contributed to the Aimsun API integration and development of the "flow" framework. Their work involved fixing API issues for OS X, merging code from the master branch, addressing coding style inconsistencies, and resolving server connection issues within the API. Further, the user made modifications and performed refactoring regarding template loading and created a script to load Aimsun templates.
autonomousreinforcement-learningvehicle-controldeep-reinforcement-learningbenchmark
google-deepmind/open_spiel
Feb 2024 - Mar 2025
OpenSpiel is a collection of environments and algorithms for research in general reinforcement learning and search/planning in games.
Role in this project:Back-end Developer
Back-end DeveloperContributions:5 PRs, 5 comments, 2 issues in 1 year 1 month
Contributions summary:Nathan primarily focused on bug fixes and improvements within the OpenSpiel game environment, specifically addressing issues related to information state representation and action handling. Their work involved modifications to the C++ and Python code, targeting areas such as the Deep Q-Network (DQN) implementation and games like Phantom Tic-Tac-Toe and Dark Hex. The changes involved adjusting parameters, correcting logic errors, and refactoring code to enhance accuracy and efficiency in the game simulations. The commits also included the removal of "magic numbers" for improved readability.
cppmultiagentgamespythondatamining
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.