Yichi Zhang
Based in New York
yichi.zhang@capitalone.com
I'm an Applied Researcher at Capital One, where I work on AI evaluation and post-training. My research focuses on designing robust automated AI evaluators, calibrated LLM predictors, and multi-AI collaboration systems, with an emphasis on improving LLM reasoning reliability, error reflection, and belief calibration.
Before joining my current position, I was a postdoctoral researcher at the Center for Discrete Mathematics and Theoretical Computer Science (DIMACS) at Rutgers University, hosted by David Pennock and Lirong Xia. I received my Ph.D. from the University of Michigan, Ann Arbor, where I was fortunate to be advised by Grant Schoenebeck. I earned my B.S. from Shanghai Jiao Tong University, China.
Recent News
- Dec 2025 I am attending NeurIPS, December 2-7, in San Diego, presenting my posters: "Evaluating LLM-contaminated Data" and "Stochastically Dominant Peer Prediction".
- Oct 2025 I am attending the 2025 INFORMS Annual Meeting, October 26-29, in Atlanta. I will present our OR revision, "A System-level Analysis of Conference Peer Review".
- July 2025 Presented our paper “Evaluating LLM-Corrupted Crowdsourcing Data Without Verification” at the Workshop on Human–Algorithm Collaboration, jointly with EC ’25.
- July 2025 We organized the 2nd Annual Workshop on Incentives in Academia (WINA), jointly with EC ’25.