Experience

  1. Stanford Research

    Stanford University

  2. Multimodal Interaction & World Model Intern

    ByteDance Seed
    • Accelerated diffusion model training by leveraging advances in representation learning.
    • Incorporated timestep scaling as a temperature factor during contrastive learning steps.
    • Improved FID from 15 to 13 by optimizing training signals and regularizers.

  3. Algorithm Intern

    Shanghai Artificial Intelligence Laboratory
    • Built optics-domain benchmarks and integrated large models (GPT-4o, LLaMA-3).
    • Evaluated mainstream LLMs on domain-specific tasks focusing on reasoning capabilities.
    • Fine-tuned models like LLaMA 3 using QLORA and PEFT for scientific Q&A.

Education

  1. Ph.D. in Operations Research, 2024 - 2029 (expected)

    Stanford University
    Advisors: Jose Blanchet, Renyuan Xu

  2. B.Sc. in Mathematics, 2020 - 2024

    Fudan University
    GPA: 3.85/4.0 (Major GPA: 3.92/4.0)
Awards
Languages