Dynamo is a cognitive robotic system built to adapt to unstructured environments and learn through interaction. By integrating foundation models with rigorous control theory, it achieves versatile, robust behavior in real-world settings.

A VLA-based pick-and-place pipeline built with LeRobot: the robot receives a natural language instruction, observes the workspace from three RGB camera views, and executes the requested vial arrangement on Jetson inference. The project studies whether randomized training data improves transfer when lighting, rack placement, and vial colors change.

The Ridgeback picks one designated person out of a multi-person scene, then fuses camera detection with point-cloud perception to follow them through clutter. Two iterations compare 2D and full 3D sensing for environment understanding.