Kai Pak

Jet Propulsion Laboratory, California Institute of Technology

Papers

Total Citations

H-Index

About

Kai Pak is a leading researcher at the intersection of autonomous robotics and artificial intelligence, with a primary focus on enabling intelligent, self-sufficient operation for planetary exploration and general-purpose robotic systems. His most impactful contribution is the development of ROSA (Robot Operating System Agent), an AI-powered interface that democratizes robotic control by allowing non-expert users to command complex systems through natural language, a work that has already garnered 11 citations since its 2025 publication. Pak is also pioneering the use of self-supervised learning to overcome critical data limitations in space robotics. His paper on self-supervised distillation for computer vision onboard planetary robots (2023, 4 citations) and his work on CLOVER, a contrastive learning framework for onboard vision (2023, 3 citations), directly address the challenges of domain shift and scarce annotated data in extraterrestrial environments. By developing models that can learn from unlabeled planetary images, Pak is laying the groundwork for the next generation of autonomous explorers capable of maximizing science return on missions beyond Mars.