Jiayi Weng is a research engineer at OpenAI working on distributed reinforcement learning infrastructure for aligning foundation models. His work focuses on high-throughput RLHF pipelines, training performance, scalability, and numerical correctness. Previously, he worked at Sea AI Lab and conducted reinforcement learning research at Tsinghua University and the Montreal Institute for Learning Algorithms. He created Tianshou and EnvPool, open-source reinforcement learning systems.
Jiayi Weng is a research engineer at OpenAI working on distributed reinforcement learning infrastructure for aligning foundation models. His work focuses on high-throughput RLHF pipelines, training performance, scalability, and numerical correctness. Previously, he worked at Sea AI Lab and conducted reinforcement learning research at Tsinghua University and the Montreal Institute for Learning Algorithms. He created Tianshou and EnvPool, open-source reinforcement learning systems.