I am currently a final year Ph.D. student at Princeton University, and I am fortunate to be advised by Prof. Sanjeev Arora. I was a research intern at AI Frontier, MSR, where I was fortunate to work with Dimitris Papailiopoulos.

I completed my undergraduate at Andrew Chi-Chih Yao’s CS pilot class at Tsinghua University, where I was advised by Prof. Wei Chen. Previously, I was also a reserach intern at Microsoft Research Asia Theory group metored by Prof. Wei Chen. In my junior year, I also visited Duke University where I had a great time working with Prof. Rong Ge.

Research Interests

I am broadly interested in the intersection of math, algorithms, and learning. My primary focus is to understand and enhance fundamental model capabilities such as reasoning and compositional generalization in large language models (LLM), and apply these general understanding to specific settings (with current emphasis on automated theorem proving). Previously I also worked on online leraning and mathematical optimization.

News

• [Aug. 2025] New paper out: Goedel-Prover-V2: Scaling Formal Theorem Proving with Scaffolded Data Synthesis and Self-Correction
• [July 2025] New paper out: AlgoTune: Can Language Models Speed Up General-Purpose Numerical Programs?
• [May. 2025] New paper out: Ineq-Comp: Benchmarking Human-Intuitive Compositional Reasoning in Automated Theorem Proving on Inequalities
• [Feb. 2025] New paper out: Unrealized Expectations: Comparing AI Methods vs Classical Algorithms for Maximum Independent Set

Contact Information

haoyu AT princeton DOT edu