Embodied AI Engineer — Robotics

Role overview An Embodied AI Engineer in Robotics develops the intelligence layer that lets physical robots perceive, reason, and act in real-world environments. The position sits at the intersection of machine learning research and hands-on robotic deployment, turning learning-based models into reliable behaviors on real machines.

Responsibilities

- Design, train, and evaluate AI models for perception, planning, and control on robotic platforms - Develop and run simulation environments to bootstrap and validate learned policies before real-world testing - Integrate learned models with robot hardware, including sensors, actuators, and middleware - Build data pipelines that capture, curate, and version multi-modal sensor data - Prototype end-to-end embodied agent behaviors and iterate from sim to hardware - Collaborate with software, controls, and mechanical teammates to ship working systems

Requirements

- Strong foundation in machine learning, deep learning, or reinforcement learning - Proficiency with at least one robotics simulation or middleware stack - Experience with one or more of: robot perception, SLAM, manipulation, locomotion, or policy learning - Solid programming skills in Python and/or C++ and familiarity with modern ML libraries - Working knowledge of robot sensor stacks and how to fuse their outputs - Track record of moving models from notebooks into deployed systems

Nice to have

- Published research or open-source contributions in embodied AI, RL, or robot learning - Experience with vision-language-action models or other foundation-model approaches for robotics - Hands-on work with real robots, including debugging on hardware and collecting real-world datasets