Osbert Bastani

Osbert Bastani is an Associate Professor at the University of Pennsylvania's Department of Computer and Information Science, where he leads the trustml@Penn research group. He is affiliated with multiple research centers including ASSET, PRECISE, PRiML, and PLClub. Previously, he completed his Ph.D. at Stanford University advised by Alex Aiken and was a postdoctoral researcher at MIT working with Armando Solar-Lezama.

His research focuses on Trustworthy Machine Learning, specifically addressing: adversarial robustness through formal verification methods, distributional robustness, uncertainty quantification via conformal prediction, fairness definitions and verification, explainability techniques (LIME/SHAP), and counterfactual explanations. His work bridges theoretical guarantees with practical applications in safety-critical systems.

He teaches graduate courses including CIS 7000: Trustworthy Machine Learning and CIS 4190/5190: Applied Machine Learning, with coursework covering robustness verification, calibrated prediction, fairness constraints, and attribution methods.

He leads the trustml@Penn research group and collaborates with multiple Penn research centers focused on embedded systems, machine learning foundations, and programming languages.