I am an assistant professor in the Industrial and Systems Engineering at King Fahd University of Petroleum and Minerals (KFUPM), Saudi Arabia. I previously served as Executive Director of the Stanford Center for AI Safety and Research Engineer at both the Stanford Intelligent Systems Lab (SISL) and Mineral-X. I obtained my Ph.D. in Mechanical Engineering from Carnegie Mellon University in the Safe AI lab.
My research integrates machine learning, optimization, and simulation to create trustworthy AI systems for high-stakes applications in both safety and sustainability domains. I focus on ensuring equitable and just energy transitions through resilient critical mineral supply chains, safety-aware geothermal and subsurface applications, and robust AI system verification and validation methods that can manage catastrophic failure risks in critical infrastructure.
Beyond research, I enjoy teaching and mentoring students at all levels. I have taught courses on engineering statistics, data-driven analytics and guided numerous research projects. I serve as Editor-in-Chief for OSCM Journal and as a reviewer for leading AI, transportation, and sustainability venues. I'm always interested in collaborating on AI safety and sustainability for critical applications and welcome discussions about potential partnerships—you can find more details in my CV or explore my publications below.
Recent Publications
- Mansur Arief, Yasmine Alonso, CJ Oshiro, William Xu, Anthony Corso, David Zhen Yin, Jef Caers, and Mykel Kochenderfer. “Managing Geological Uncertainty in Critical Mineral Supply Chains: A POMDP Approach with Application to US Lithium Resources.” arXiv preprint. [Link]
- Liam Kruse, Alexandros Tikas, Harrison Delecki, Mansur Arief, Mykel Kochenderfer. "Enhanced Importance Sampling through Latent Space Exploration in Normalizing Flows." arXiv preprint. [Link]
- Arief, Mansur, Mike Timmerman, Jiachen Li, David Isele, Mykel Kochenderfer. "Optimized Guided Meta Training for Intelligent Agents under Highly Interactive Driving Scenarios." IEEE RA-L. [Link]
