Zhengxing Chen
Research Scientist at Meta
Menlo Park, California, United States
Summary
🤩
Rockstar🎓
Top SchoolZhengxing Chen is a research scientist with 11 years of experience specializing in applied reinforcement learning and large-scale user modeling, currently working on Llama 3 RLHF research at Meta in Menlo Park. He has driven industry-scale RL projects across ranking, retrieval, and value optimization, and was a main contributor to Meta’s open-source ReAgent platform where he implemented and evaluated RL models like DQN and feature-importance tools. His work spans both foundational research—co-authoring multiple WWW and workshop papers on scaling user representations, AutoML, and RL systems—and practical product impact in ads and notifications. With a PhD in Computer Science and internships at Google, Instagram, and EA, he blends rigorous academic training with hands-on production engineering. A less obvious strength is his cross-domain fluency, from game matchmaking analysis early in his career to state-of-the-art generative-model RL research today.
11 years of coding experience
7 years of employment as a software developer
Bachelor's degree Computer and Information Science, Bachelor's degree Computer and Information Science at Beijing University of Posts and TelecommunicationsDoctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at Northeastern University
Github Skills (10)
pytorch10
dqn10
machine-learning10
deep-learning10
python10
reinforcement-learning10
modeling9
feature-engineering9
trainings9
data-analysis8
Programming languages (2)
Github contributions (5)
facebookresearch/ReAgent
Oct 2018 - Aug 2022
A platform for Reasoning systems (Reinforcement Learning, Contextual Bandits, etc.)
Role in this project:
ML Engineer
ML EngineerContributions:233 commits, 90 PRs, 9 pushes in 3 years 10 months
Contributions summary:Zhengxing's commits primarily focus on implementing and evaluating RL models within the Reinforcement Learning platform. Specifically, the commits introduce and refactor code related to training and evaluating the performance of various RL models, like Deep Q-Networks (DQN). The commits also include the development of functionality for feature importance analysis. The contributions involve defining model structures, establishing training workflows, and creating tools for assessing model performance in the context of reinforcement learning tasks.
reinforcement-learningcontextualbanditscontextual-banditsreinforcement
czxttkl/czxttkl.github.io
Jan 2015 - Jun 2019
Contributions:34 pushes in 4 years 6 months
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.