I am a CS PhD student at Yale University, advised by Professor Arman Cohan. I was very fortunate to have the late Professor Dragomir Radev as my advisor during my first two years at Yale. Before coming to Yale, I obtained my B.Eng. from Zhejiang University with a dual B.S. from UIUC. Like many engineers who entered the field in the era of LLMs, my research interests have evolved alongside the rapid advancement of AI. Currently, I focus on (1) building foundation models that deliver expert-level reasoning, multimodal understanding, and knowledge-intensive capabilities for high-stakes domains such as science, (2) developing scalable, expert- and community-driven evaluation and alignment methods that make expert-domain model behavior measurable, comparable, and improvable, and (3) quantifying and mitigating societal and user-facing risks in real workflows, and designing human-AI systems that meet real user needs for responsible deployment.

About Collaboration

I am fortunate to work closely with many collaborators whose insight and generosity I deeply value. In my own collaborations, I value thoughtfulness, fairness, and substantive contribution. In particular, I take first authorship only when I have played a leading role and contributed at least as much as any other author, and I join projects as a coauthor only when I can make a meaningful contribution. If you are interested in any of the above directions or my past research, I would be very happy to connect!

I dedicate 2×30 minutes each week to offering guidance, suggestions, and mentorship. If you would like to talk about life, career paths, graduate school applications, or NLP/LLM research ideas, feel free to drop me an email :)

I aim to make my research as reproducible as possible. For papers on which I am a core author or corresponding author, if you encounter any issues in reproduction, please contact me and I will do my best to help.

Workshop & Tutorial Organization

• CVPR 2026, ICCV 2025 Workshop on Knowledge-Intensive Multimodal Reasoning
• AAAI 2026 Workshop on New Frontiers in Information Retrieval
• AAAI 2026 Workshop on AI for Scientific Research
• NeurIPS 2025 Workshop on Socially Responsible and Trustworthy Foundation Models (ResponsibleFM)
• CVPR 2025 Tutorial on Multimodal Mathematical Reasoning
• New England NLP (NENLP) Meeting Series 2025
• ICLR 2025 Workshop on LLM Reasoning and Planning

Program Committee

• Area Chair: NeurIPS 2026, ACL Rolling Review 2026, NeurIPS 2025
• Reviewer: ACL Rolling Review 2021–25 and other ACL events, CVPR, COLM, ICLR, NeurIPS, ICML