Me
Wanjia Zhao (赵万佳)
CS PHD Student @ Stanford

About

Hi, I am a second-year Computer Science Ph.D. student at Stanford, where I am fortunately to be advised Prof. James Zou and Prof. Ludwig Schmidt. I obtained my B.S. (Honors) degree in Mathematics and Applied Mathematics from Zhejiang University.
Previously, I was a research intern at Microsoft Research and DeepSeek AI. Before that, I was also a visiting research student at UCLA Scalable Analytics Institute (ScAi), supervised by Prof. Yizhou Sun and Prof. Wei Wang.

My research focuses on LLM reasoning and generative models for science, with applications in mathematics, physics, and scientific discovery.

Education

Ph.D.           Sep. 2024 - Present
                     Stanford University, Palo Alto, CA, U.S.
                     Ph.D. student in Computer Science
                     Advisor: Prof. James Zou and Prof. Ludwig Schmidt
B.S.              Sep. 2020 - Jun. 2024
                     Zhejiang University (ZJU) , Hangzhou, China.
                     B.S.(Honors) in Mathematics and Applied Mathematics
                     Advisor: Prof. Fei Wu

Experiences

  • Jun. 2025 -- Sept. 2025, AI Frontiers, Microsoft Research, Redmond, WA, US
         Research Intern, Work on RL and Tool Use in LLM Reasoning
  • Jun. 2024 -- Sept. 2024, DeepSeek AI , Beijing, China
         Research Intern, Work on LLM for Mathematical Reasoning
  • Feb. 2024 -- Jun. 2024, Machine Learning Group, Microsoft Research Lab - Asia, Beijing, China
         Research Intern, Work on LLM for decision making
  • Jun. 2023 -- Oct. 2023, University of California Los Angeles (UCLA), Los Angeles, CA, U.S.
         Visiting Research Student, Work on Physics informed ML
         Supervisor: Prof. Yizhou Sun and Prof. Wei Wang
  • News

    Selected Publications and Manuscripts, (Full List: here)

    PRISM-Physics: Causal DAG-Based Process Evaluation for Physics Reasoning
    Preprint 2025

    GeoAda: Efficiently Finetune Geometric Diffusion Models with Equivariant Adapters
    NeurIPS 2025
    [ Paper] [www] [Code]

    SiriuS: Self-improving Multi-agent Systems via Bootstrapped Reasoning
    NeurIPS 2025
    [ Paper] [www] [Code]

    OpenThoughts: Data Recipes for Reasoning Models
    Preprint, 2025

    Protein Large Language Models: A Comprehensive Survey
    Findings of EMNLP 2025
    [ Paper] [www] [Github]

    Physics-Informed Regularization for Domain-Agnostic Dynamical System Modeling
    NeurIPS, 2024
    Best Paper Award @ NeurIPS 2023 DLDE

    Deepseek-prover-v2: Advancing formal mathematical reasoning via reinforcement learning for subgoal decomposition
    Preprint, 2025

    DeepSeek-prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search
    ICLR 2025

    Selected Honors & Awards

    • NeurlPS 2024 Scholar Award, 2024
    • Outstanding Graduates of Zhejiang Province and Zhejiang University, 2024
    • SenseTime Scholarship, 2023
    • Best Paper Award, NeurIPS 2023 Workshop (Deep Learning and Differential Equation), 2023
    • Zhejiang University Scholarship - First Prize (Top 3%), 2023
    • CKC Honors College Scientific Research Scholarship (Top 10 Projects), 2023
    • Silver Medal of Chinese Physics Olympiad, 2019
    • First Prize of Chinese Physics Olympiad, 2019

    Academic Services

    Conference Reviewer: NeurIPS, ICLR, ICML
    Volunteer: KDD 2023, NeurIPS 2023

    Miscellaneous

    I love running, piano, debating, traveling and so much more.
    🎤 Champion and Best Debater of the CKC Honors College Debate Tournament
    🏃‍♀️ Runner-up of the Women's Pentathlon at Zhejiang University
    🎹 Pianist of amateur level 10

    Top