I’m a PhD candidate at the Institute of Computing Technology, Chinese Academy of Sciences (State Key Lab of AI Safety), currently working as a research intern at Tencent (LLM Department / T3), building computer-using agents and verifiable RLVR training loops for Tencent Hy. Before that, I built language agents at SkyworkAI, Apple, and MSRA. My research interests span agentic reinforcement learning and verifiable post-training, long-context memory mechanisms of LLMs, and AI safety and trustworthy alignment. Highlights: ACL 2026 (Oral), 1500+ citations. Beyond academia, I’m deeply passionate about photography, particularly conceptual, street, new topographic, and event photography. When I’m not behind the camera or immersed in research, you can find me speeding on the racetrack as an amateur race car driver or delving into games as an avid Steam user.
— still, as the old banner said: a (boyfriend / racer / student / Steam user).
GitHub · Google Scholar · CV (English, PDF) · 简历 (中文, PDF) · 履歴書 (日本語, PDF) · Text version of this page