About Me
Hi! I'm Chengjiao Ye, a master's student at AI4GC Lab, Zhejiang University, advised by Shengyu Zhang. My research focuses on confidence calibration and uncertainty estimation for large language models and AI agents, especially how to make agentic systems better understand when their answers and reasoning trajectories are reliable.
I am interested in building more trustworthy, interpretable, and robust LLM/agent systems. Open to discussions and collaborations on LLM calibration, agent evaluation, and trustworthy AI.
Research Interests
- LLM Calibration - Estimating and improving the reliability of model confidence in open-ended generation and reasoning tasks.
- Agentic Confidence Calibration - Calibrating confidence at the trajectory level for LLM-based agents rather than only at the final-answer level.
- Trustworthy AI Evaluation - Designing evaluation protocols and metrics for reliability, uncertainty, and robustness in LLMs and AI agents.
Education
- 2026 - present · M.S. in Artificial Intelligence, Zhejiang University
- 2022 - 2026 · B.S. in Automation, Wuhan University