Menlo Park, United States
Zheqing (Bill) is the engineering manager for the Applied Reinforcement Learning group within Meta AI. The Applied Reinforcement Learning group aims to bring the state-of-the-art reinforcement learning technology to life by pushing the frontier of industry-ready reinforcement learning technologies. The team owns Meta's reinforcement learning stack and spans across dozens of product reinforcement learning usecases. Our application and research areas include but are not limited to recommender system, user contextual understanding, control mechanisms, and many more. The team also developed and currently owns Meta's open source reinforcement learning platform ReAgent (reagent.ai).
Bill's personal research interest lies in bridging the gap between theoretical reinforcement learning with practical applications. Some of the reinforcement learning topics that Bill is particularly interested in are improving sample complexity, handling nonstationarity, generalization of value functions, and representation of agent-environment history.