Yitong Lin

Hello! 👋🏻 I am a Master's student at Zhejiang University, advised by Prof. Xiaoxue Ren. I work on AI Agents, Agentic Evaluation, and Reinforcement Learning.

My research explores how AI agents can evaluate their actions and use that evidence to improve. I am particularly interested in agentic evaluation, reward signals for reinforcement learning, and self-improvement through interaction. My recent work studies how execution and evaluation can co-evolve at test time, while my earlier work on Code LLMs examines API evolution and executable benchmarks for code generation.