Jens Kober

Jens Kober is an Associate Professor in the Cognitive Robotics department at Delft University of Technology's Faculty of Electrical Engineering, Mathematics and Computer Science, and a key member of the TU Delft Robotics Institute. His research bridges machine learning and robotics for real-world applications.

His research focuses on Motor Skill Learning, Reinforcement Learning, and Interactive Learning for robotics, with applications spanning robot manipulation, locomotion, surgical assistance, and wind turbine control. Key areas include:

• Developing robust learning algorithms for compliant robots
• Creating interactive learning frameworks for human-robot collaboration
• Applying deep learning to robot control under uncertainty
• Transferring machine learning techniques to medical robotics

His recent publications reveal strong trends in interactive imitation learning and reinforcement learning for dynamic tasks, particularly in robot manipulation (e.g., pick-and-place ambiguity resolution) and locomotion (explosive jumping with soft quadrupeds). The work increasingly integrates vision systems and focuses on real-world applicability in retail, dentistry, and renewable energy.

Scientific recognition includes:

• Robotics: Science and Systems Early Career Award (2022)
• IEEE-RAS Early Academic Career Award (2018)
• Georges Giralt PhD Award for best European Robotics thesis (2013)

Kober actively supervises PhD students and leads major collaborative projects including the NWO-funded DL-foRCe (Deep Learning for Robust Robot Control), Robot-Assisted Tooth Removal, and the industry-backed AI for Retail Lab with Ahold Delhaize. His work connects academic research with industrial applications through grants from NWO, TKI Wind op Zee, and industry partnerships.

He directs the Cognitive Robotics group within the TU Delft Robotics Institute, collaborating closely with the Delft Center for Systems and Control and the ELLIS Unit Delft. His team develops foundational learning algorithms while deploying them in applied settings like the AIR Lab Delft test site for retail robotics.