PhD student, Arizona State University
1 paper at NeurIPS 2025
This paper argues that post-deployment monitoring in clinical AI is underdeveloped and proposes statistically valid and label-efficient testing frameworks to ensure reliability and safety.