I have deep expertise in understanding and improving the reliability of AI systems. My work focuses on developing AI that inherently “knows what it does and does not know,” enabling adaptive computation and more efficient decision-making. Ultimately, my goal is to optimize performance in long-tail scenarios while maintaining robustness under extreme uncertainty.
I completed my PhD at MIT with a thesis on Reliable Foundation Models: From Instance-Level Assessment to Uncertainty-Aware Adaptation. Prior to MIT, I deployed production-scale ML systems at Meta Instagram and NAVER AI Lab. I am joining Tesla as a Senior Autopilot Machine Learning Engineer, focusing on the next generation of reliable end-to-end autonomous systems.
Full publication list available at Google Scholar