Yuandong Tian
Co-Founder at Stealth AI Startup
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolYuandong Tian is a Research Scientist and Senior Manager (now Research Director) at Meta AI Research (FAIR) in Menlo Park with 12 years of experience combining deep learning research and engineering. He leads work on reinforcement learning and the theoretical and empirical understanding of self-supervised deep networks, driving ML/RL-guided optimization toward deployable systems. His hands-on open-source contributions to prominent FAIR projects—improving the DarkForest Go engine’s CNN player and training pipeline and refactoring the ELF game-research platform—demonstrate a rare balance of research depth and production-grade system design. Earlier roles include software engineering at Google X on self-driving cars and a Ph.D. in Robotics from Carnegie Mellon, grounding his research in control and embodied AI. He also holds JLPT Level 2 Japanese proficiency, a useful asset for international collaboration.
13 years of coding experience
12 years of employment as a software developer
Ph.D Robotics Institute, Ph.D Robotics Institute at Carnegie Mellon University
Github Skills (20)
c-language10
convolutional-neural-networks10
machine-learning10
refactor10
lua10
neural-network10
refactoring10
game-development10
cprogramming-language10
ai9
header-files9
modeling8
trainings8
algorithm7
algorithms7
Programming languages (4)
Github contributions (5)
facebookresearch/darkforestGo
Jun 2016 - Nov 2016
DarkForest, the Facebook Go engine.
Role in this project:
ML Engineer
ML EngineerContributions:24 commits, 2 PRs, 36 pushes in 5 months
Contributions summary:Yuandong primarily focused on improving the engine's core functionality within the DarkForest Go project. They modified various Lua scripts related to CNN player implementation (cnnPlayerV2/cnnPlayerMCTSV2.lua), fixing issues and adjusting parameters. Further contributions included updates to the training script and configuration files, along with the removal of legacy files and updates to documentation (README). This indicates an effort to enhance the model's performance and integration within the Go engine.
golangfacebook
facebookresearch/ELF
Jul 2017 - Jun 2018
An End-To-End, Lightweight and Flexible Platform for Game Research
Role in this project:Back-end Developer
Back-end DeveloperContributions:176 commits, 14 PRs, 328 pushes in 11 months
Contributions summary:Yuandong primarily refactored and updated core classes, specifically altering class inheritance and dependencies within the game research platform. The changes involved refactoring code related to AI, including renaming classes and updating header files. They also removed a dependency on rlmethod_forward. These contributions indicate a focus on improving the platform's structure and dependencies.
flexiblegame-engineend-to-endgame-development
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.