Chengrui Qu

Chengrui Qu

PhD Student at Caltech

About

I am a first-year PhD student in the Caltech Rigorous Systems Research Group, advised by Prof. Adam Wierman and Prof. Eric Mazumdar at the Computing + Mathematical Sciences (CMS) Department, California Institute of Technology. Most people call me Ray. Before joining Caltech, I obtained my B.Sc. degree at Peking University.

Research

My research develops theoretical and algorithmic foundations for multi-agent alignment, ensuring AI agents, including LLMs, can reliably cooperate with each other and with people. I approach this through the lens of game theory and robust decision-making, building agents that generalize to novel partners and environments rather than overfitting to training-time conventions. I’m always happy to connect; feel free to reach out if you’d like to discuss research, collaborations, or entrepreneurial opportunities.

News

Sep 2026
Our paper “Training Generalizable Collaborative Agents via Strategic Risk Aversion” has been accepted as an oral presentation at NeurIPS 2026 (top 0.3% of submitted papers)! Many thanks to all my wonderful collaborators!
Jul 2026
Our paper “Behavioral Game Theory for Collaborative LLM Agents” has been accepted as a spotlight oral at the COLM 2026 Workshop on Agent Behavior! Many thanks to all my wonderful collaborators!
Jun 2026
Excited to be joining the Core AI team at IBM and the Red Hat AI Innovation team as a Research Intern, advised by Akash Srivastava!
Jan 2026
Our work “Distributionally Robust Cooperative Multi-agent Reinforcement Learning with Value Factorization” got accepted by ICLR 2026! Many Thanks to all my wonderful collaborators!
Dec 2025
Happy to share our work on Decision-Dependent Distributionally Robust Optimization (DD-DRO) in Rio at CDC 2025! This work is a first step toward generalizing DRO to endogenous uncertainty, with provable guarantees.

Latest Posts

Aug 2026
Knowledge-Centric Self-Improvement
Aug 2026
The Submartingale Condition: When Self-Improvement Actually Improves

Selected Publications

  1. settings.png
    Hybrid Transfer Reinforcement Learning: Provable Sample Efficiency From Shifted-dynamics Data
    Chengrui Qu, Laixi Shi, Kishan Panaganti, Pengcheng You, and Adam Wierman
    AISTATS (oral, top 2%), 2025
    arXiv
  2. dd_dro.png
    Decision-Dependent Distributionally Robust Optimization with Application to Dynamic Pricing
    Chengrui Qu, Huiwen Jia, and Pengcheng You
    IEEE CDC, 2025
    arXiv
  3. spidr.gif
    SPiDR: A Simple Approach for Zero-Shot Safety in Sim-to-Real Transfer
    Yarden As, Chengrui Qu, Benjamin Unger, Dongho Kang, Max Hart, Laixi Shi, Stelian Coros, Adam Wierman, and Andreas Krause
    NeurIPS, 2025
    arXiv
  4. robust_vdn.png
    Distributionally Robust Cooperative Multi-Agent Reinforcement Learning via Robust Value Factorization
    Chengrui Qu, Kishan Panaganti, Christopher Yeh, and Adam Wierman
    ICLR, 2026
    arXiv
  5. mpe_play.gif
    Training Generalizable Collaborative Agents via Strategic Risk Aversion
    Chengrui Qu, Yizhou Zhang, Nicolas Lanzetti, and Eric Mazumdar
    NeurIPS (oral, top 0.3%), 2026
    arXiv
  6. diversity.png
    Understanding Agent Scaling in LLM-Based Multi-Agent Systems via Diversity
    Yingxuan Yang, Chengrui Qu, Muning Wen, Laixi Shi, Ying Wen, Weinan Zhang, Adam Wierman, and Shangding Gu
    arXiv preprint, 2026
    arXiv