Two papers accepted at NeurIPS 2025, one at ASE 2025
One paper accepted at ACL 2025, one at KDD ADS 2025
Two papers accepted at ICLR 2025, one at NAACL 2025 (UFO paper)
Released work on Large Action Models, covered by media outlets such as Jiqizhixin and XinZhiYuan
Published a comprehensive 90+ page survey on LLM-powered GUI agents covering 500+ papers, reported by Jiqizhixin
Led development of UFO²: a next-generation AgentOS for Windows desktop with HostAgent/AppAgents architecture, hybrid control detection, and cross-application orchestration, widely covered in media
Background
Principal Researcher in the Data, Knowledge and Intelligence (DKI) group at Microsoft
Current research focuses on GUI Agents and Computer-Using Agents (CUAs) powered by large language models to enhance user experiences on computer systems
Previously worked on AIOps projects using both traditional and LLM-based approaches to drive innovation and technology transfer at Microsoft
Core developer and maintainer of UFO, the first GUI agent for Windows OS
Maintains a public survey on large language model-powered GUI agents