Published several papers, including 'VideoScore2: Think before You Score in Generative Video Evaluation' and 'VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use'.
Research Experience
Currently interning at NVIDIA ADLR. Involved in research projects at TIGER-Lab.
Education
Currently a second-year CS Ph.D. student at the University of Waterloo, advised by Prof. Wenhu Chen. Received a bachelor's degree in Computer Science from Zhejiang University.
Background
Research interests include LLM/VLM post-training, including alignment, evaluation, and applications. Particularly interested in reinforcement learning for reasoning tasks and how they can better use tools to solve problems.
Miscellany
Personal project includes VerlTool, an initial exploration of using tools for reinforcement learning with a neat and easy-to-use codebase.