Hangyu Mao

Research Professor, Ph.D. Supervisor, and Deputy Director of the AI Center, Institute of Microelectronics, Chinese Academy of Sciences

Hangyu Mao is a Research Professor, Ph.D. Supervisor, and Deputy Director of the AI Center at the Institute of Microelectronics, Chinese Academy of Sciences. His research spans reinforcement learning, large language models, agentic systems, and AI chip architectures. He has published more than 50 papers and introduced new training paradigms for LLM-based agents, including ARPO (Agentic Reinforced Policy Optimization) and TPTU (Task Planning and Tool Usage), which enhance agents’ capabilities in complex reasoning and tool use. His work has received more than 2,500 Google Scholar citations over the past five years. Mao has led multiple national-, provincial-, and ministerial-level research projects, securing total funding exceeding RMB 100 million. He and his team have received recognition including a representative case award for large-model applications at the Global Digital Economy Conference, first place in a reinforcement learning competition at NeurIPS, and the Outstanding Doctoral Dissertation Award in Multi-Agent Systems from the China Computer Federation (CCF). He has also served as the head of R&D teams at several leading technology companies, where he led the development of agentic products that surpassed 10 million users and 10 million monthly active users, achieving large-scale real-world deployment. He is committed to advancing end-to-end software–hardware co-design across the full “application–algorithm–system–chip” stack, driving the creation of real-world value through AI agents.