I'm an AI researcher working on principled approaches to uncertainty, risk, and robust decision-making in complex real-world systems. Currently, I'm the chief scientist at Valgo, where I focus on risk quantification for physical AI. I am also an AI/ML scientist at Terra AI, developing uncertainty quantification methods for large-scale sustainable resource projects, including mineral exploration and carbon sequestration. Previously, I earned my PhD at Stanford University, where my research focused on sampling and estimation for safety validation of autonomous systems like self-driving cars and autonomous aircraft.
At Stanford, I was a member of the Stanford Intelligent Systems Laboratory (SISL) and the Stanford Center for AI Safety, advised by Mykel Kochenderfer. I collaborated with sponsors including NASA JPL, Ford, Motional, and Nissan. Before Stanford, I earned my Bachelor's degree from Georgia Tech, where I worked on attitude determination and control for small satellites.
Publications
Failure Probability Estimation for Black-Box Autonomous Systems using State-Dependent Importance Sampling Proposals
Harrison Delecki, Sydney M. Katz, and Mykel J. Kochenderfer
International Conference on Control, Decision and Information Technologies (CoDIT) 2025
Diffusion Failure Sampling for Evaluating Safety-Critical Autonomous Systems
Harrison Delecki, Marc R. Schlichting, Mansur Arief, Anthony Corso, Marcell Vazquez-Chanlatte, and Mykel J. Kochenderfer
International Conference on Engineering Reliable Autonomous Systems (ERAS) 2025
Entropy-regularized Point-based Value Iteration
Harrison Delecki, Marcell Vazquez-Chanlatte, Esen Yel, Kyle Wray, Tomer Arnon, Stefan Witwicki, and Mykel J. Kochenderfer
International Conference on Control, Decision and Information Technologies (CoDIT) 2025
Deep Normalizing Flows for State Estimation
Harrison Delecki, Liam A. Kruse, Marc R. Schlichting, and Mykel J. Kochenderfer
International Conference on Information Fusion 2023
Model-based Validation as Probabilistic Inference
Harrison Delecki, Anthony Corso, and Mykel J. Kochenderfer
Learning for Dynamics and Control Conference 2023, Oral
How Do We Fail? Stress Testing Perception in Autonomous Vehicles
Harrison Delecki, Masha Itkina, Bernard Lange, Ransalu Senanayake, and Mykel J. Kochenderfer
International Conference on Intelligent Robots and Systems 2022