Summer Yue
Director, Alignment at Meta
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolSummer Yue is a research and engineering leader with 11 years of experience building RL, LLM, and large-scale ML systems across startups and Google/Meta ecosystems. She has led teams shipping production RLHF for Bard, pushed reproducible RL infrastructure and robotics work at Google Brain, and ran a 64-person ML research org at Scale focusing on synthetic data, agent reasoning, and safety evaluations. Currently Director of Alignment at Meta, she blends hands-on contributions (e.g., TF-Agents testing and PPO work) with strategy for superintelligence alignment and scalable oversight. Known for translating frontier research into reliable production pipelines, she’s equally comfortable debugging unit tests as she is setting research agendas to mitigate reward hacking and other systemic risks. Based in San Francisco, she prioritizes practical, safety-first approaches that maximize AI’s positive impact.
11 years of coding experience
4 years of employment as a software developer
Bachelor of Applied Science - BASc Computer Science, Bachelor of Applied Science - BASc Computer Science at Jerome Fisher M&T ProgramThe Wharton School
English, Chinese
Github Skills (7)
mle10
ppp10
tensorflow10
python10
reinforcement-learning10
ml10
testing10
Programming languages (3)
Github contributions (5)
tensorflow/agents
Feb 2020 - Dec 2021
TF-Agents: A reliable, scalable and easy to use TensorFlow library for Contextual Bandits and Reinforcement Learning.
Role in this project:
ML Engineer
ML EngineerContributions:2 reviews, 96 commits, 1 PR in 1 year 10 months
Contributions summary:Summer primarily contributed to improving the unit tests for the reinforcement learning agents within the TensorFlow Agents library. Their work included fixing a bug in a unit test related to actor distribution networks, enhancing the effectiveness of behavioral cloning agent tests by replacing random trajectories, and creating and modifying PPO (Proximal Policy Optimization) agents. The user also implemented an option to clip value predictions and worked on updating documentation related to policy information in the trajectory files.
scalabletf-agentsmultiagent-reinforcement-learningtensorflow-librarymulti-armed-bandits
summer-yue/alphago0
Nov 2017 - Mar 2020
Contributions:125 commits, 18 PRs, 112 pushes in 2 years 4 months
alphago-zerogamezeroalphago
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.