Dexterous, contact-rich manipulation
How robots can learn precise manipulation skills when success depends on force, contact, timing, and object geometry.
AI & Robotics PhD researcher
Sami Osborn
Robot Learning LabImperial College London
Questions & directions
Research focus
Problems at the boundary of learning, perception, control, and physical systems.
Imitation learning and generalisation
How robots can acquire new behaviours from demonstrations without requiring extensive task-specific retraining.
Representations and perception for manipulation
How representations of objects, scenes, and physical interaction can support robust manipulation decisions.
Reinforcement learning and control
How learning-based methods and explicit control structure can work together in reliable embodied systems.
Things built
Selected projects
SO100 Arm
Manipulation project exploring kinematics, calibration, and reliable arm control primitives.
BalanceBot
Two-wheel self-balancing robot built around an ESP32, with a focus on sensing, embedded control, and disciplined hardware bring-up.
MonoSLAM
Monocular SLAM implementation focused on geometric perception, robust bootstrapping, and frontend failure modes.
Double DQN
Reinforcement learning project focused on overestimation bias and stable value learning.
Notes & logs
Recent writing
Background
Research, engineering, and technology in practice
Before robotics research, I invested in deep technology at Bootstrap Europe and worked in climate-technology investment banking at Barclays. That experience informs how I think about moving research prototypes into deployed systems.
About and background