Benjamin is a Postdoctoral Researcher within the Mobile Robotics Group (MRG), where he focuses on making robots that can be robustly deployed in diverse, novel environments.
He originally joined the Oxford Robotics Institute in 2020 as a DPhil student at Wolfson College, supervised by Prof. Paul Newman as part of the Autonomous Intelligent Machines and Systems (AIMS) CDT. His doctoral research explored how general representations learned by vision foundation models could be used to represent and compare image datasets across domains and tasks. He applied these solutions to downstream robotics tasks, such as place recognition and selecting relevant maps for robot localisation.
Prior to his DPhil, Benjamin earned a BSc and MSc in Microengineering from EPFL, where he worked on human motion modelling under Prof. Aude Billard. He also holds an MSc in Aerospace Engineering from Polytechnique Montréal, where he researched multi-robot SLAM and object pose estimation under Prof. Giovanni Beltrame.
During his studies, Benjamin was a visiting student in Prof. Mac Schwager's group at Stanford University and interned with NASA JPL’s CoSTAR team. There, he applied his work on object pose estimation to localise objects in underground environments for the DARPA Subterranean Challenge.
Most Recent Publications
Diversity Without Labels: Task-Agnostic Coreset Selection Across Perception and Action
Diversity Without Labels: Task-Agnostic Coreset Selection Across Perception and Action
Beyond Visual Quality: A Study of Test-Time Planning with World Action Models
Beyond Visual Quality: A Study of Test-Time Planning with World Action Models
InfraVLA: Extending Vision-Language-Action Navigation with Infrastructure Cameras
InfraVLA: Extending Vision-Language-Action Navigation with Infrastructure Cameras
Fantastic features and where to find them: a probing method to combine features from multiple foundation models
Fantastic features and where to find them: a probing method to combine features from multiple foundation models
Multi-Robot Planning and Control from CCTV Camera Networks in a Real Warehouse
Multi-Robot Planning and Control from CCTV Camera Networks in a Real Warehouse