Karl Pertsch

Karl Pertsch is a member of the technical staff at Physical Intelligence, where he works on training robot foundation models. He previously served as a postdoctoral researcher at UC Berkeley and Stanford University, co-advised by Sergey Levine and Chelsea Finn, and completed his PhD in computer science at the University of Southern California. He is best known for co-leading the Open X-Embodiment project and co-creating the DROID dataset, a large-scale in-the-wild robot manipulation dataset. His research focuses on building diverse robot datasets, training large-scale robot policies, and developing scalable evaluation methods for robot foundation models.