What If Robots Could Learn to Work Underwater? ULOHA is an underwater bimanual robot learning platform built for studying robot learning in a very different environment. It combines custom leader-follower hardware with LeRobot, multi-view sensing, teleoperation, demonstration collection, policy training and autonomous deployment. The team tested ACT, Diffusion Policy and SmolVLA on real underwater tasks including inter-arm transfer, shared-object manipulation and buoyancy-driven interception. What makes it interesting is that underwater robotics adds problems you don't normally see on a workbench, including buoyancy, bubbles, moving objects and different dynamics. A fascinating platform for exploring how robot learning changes when the environment itself fights back.