Somil Bansal

Stanford AeroAstro’s Safe and Intelligent Autonomy Lab (SIA Lab) is recruiting two postdoctoral researchers for projects at the intersection of robotics, machine learning, control, and safety. The first position focuses on the safety of learning-enabled autonomous systems through a NASA University Leadership Initiative project involving academic, industry, and government collaborators. The research addresses design-time safety assurances and safety-aware learning, runtime safety monitoring and control, simulation-based testing and validation, and post-deployment learning and adaptation. Although the primary applications are in aviation, the methods can extend to robotics and other autonomous systems. The second position explores reliable foundation models and evaluation for robot autonomy. Potential research directions include steering and controlling foundation models, developing reliable long-horizon world models, and enabling robust planning and decision-making under model uncertainty and errors. Researchers may also develop methods to evaluate robot policies, identify failure modes, and systematically stress-test autonomous systems, particularly through simulation-based approaches. The positions are intentionally broad, giving postdoctoral researchers flexibility to shape projects around their backgrounds and interests within the SIA Lab’s research environment at Stanford University.