Research Scientist - Robot Learning (VLA / WAM)

Spaitial

London, United Kingdom · Robot Learning & AI

Lead end-to-end training of vision-language-action and world-action policies, from data through to a policy running on a physical robot. The work includes adapting vision-language backbones for control and closing the sim-to-real gap using domain randomisation and calibration. Senior candidates should have already trained working manipulation policies and be able to explain why others failed.