Controls Engineer, Locomotion

About this role

GRAM is seeking a Controls Engineer to develop and deploy locomotion policies for multi-legged robots that can traverse complex 3D structures in any orientation. You'll own the complete stack from simulation through hardware, combining reinforcement learning, contact mechanics, and real-world debugging to enable robots that climb, invert, and navigate arbitrary geometry without centralized control.

What you'll do

• Own end-to-end locomotion policy development from simulation environment design through physical robot deployment
• Build contact-aware RL environments with domain randomization for diverse geometries, contact mechanics, and body orientations
• Develop vision-conditioned foothold selection and path planning systems using raw sensor data and local geometry
• Collaborate with mechatronics and adhesion teams to integrate controllers with actual contact mechanisms
• Create sim-to-real evaluation frameworks testing coverage, recovery, failure modes, and hardware robustness
• Extend locomotion capabilities toward multi-robot coordination and collaborative structure traversal

What they're looking for

• Robot locomotion stack development on physical hardware
• Reinforcement learning and policy deployment
• Multi-legged or contact-rich platform experience (hexapods, quadrupeds, climbing robots)
• Python and modern robotics frameworks (Isaac Lab, legged_gym, MuJoCo, Drake, ROS2)
• Contact mechanics and classical control theory
• Sim-to-real transfer and cross-abstraction debugging
• Vision-conditioned control and perception integration
• Whole-body contact-aware control (TSID, Pinocchio, OCS2)