Come and Build a Silicon Dream with Me

About me

👋 Hi, I'm Yingchao Li/Yanchuan (李迎超/延川), currently serving as Research Lead at MetaGPT. Prior to this, I worked as a Post-Training Engineer with Anthropic Engineer Team. I specialize in developing efficient, reliable, and safe multi-agent systems (MAS) and contribute to the infrastructure development of MAS, such as learning capabilities, reward design and memory systems. Before diving into the agent field, I also worked on large-scale AI Infra and AI Platform projects at Alibaba Cloud and HighFlyer AI. I have a deep understanding of both the software and hardware engineering of modern AI systems, and I can provide technical support throughout the entire lifecycle of AI system and application development.

🦾 In parallel with industry development, I am progressively delving deeper into academic research, conducting studies in multi-agent systems, reinforcement learning, imitation learning, transfer learning, and memory systems. My vision is to make AI-driven embodied or disembodied agents become important partners and key facilitators in human production and life, helping humans to drive all programmable entities through natural language, and contributing to the universality and accessibility of AI.

Education

University of Washington - Seattle

MS in CFRM, Sep 2022 - Dec 2023

Research Interest

Multi‑Agent Systems (MAS)
Agentic LLM Architectures
Learning and Memory System
Reinforcement Learning and Reward Design

Publications

‡: Equal Contribution

†: Coresponding Author

Guo Y., Xu Z., Yao Z., Lu Y., Lin J., Hu S., Tang Z., Li Y.†, Wang H.†, Chen R.† (2026). Octopus: Agentic Multimodal Reasoning with Six-Capability Orchestration, https://arxiv.org/abs/2511.15351, submitted to CVPR 2026

Zhang J., Peng Y., Kong F., Wu Y., Yu Z., Xiang J., Ruan J., Li Y., Song M., Tang X., Liu B., Wu C., Luo Y. (2026). Automating Environments For Measuring Agentic Learning, submitted to ICLR 2026

Yu Z., Zhang J., Su H., Zhao Y., Wu Y., Deng M., Xiang J., Tang L., Li Y, Luo Y., Liu B., Wu C. (2026). ReCode: Unify Plan and Action for Universal Granularity Control, https://arxiv.org/abs/2510.23564, submitted to ICLR 2026

Li Y., (2025). Human-Centered Editable Speech-to-Sign-Language Generation via Streaming Conformer-Transformer and Resampling Hook, https://arxiv.org/abs/2506.14677, CoRR 2025

Agrawal D.‡, Boussioux L.‡, Li Y.‡ , Wang Y. (2024). LLM Reasoning Patterns in Complex Programming and Optimization Challenges, Presented at AAAI 2024 Bridge Program – Combining AI and ORMS for Better Trustworthy Decision Making

Page updated

Google Sites

Report abuse