Research Engineer (ML, Perception and Manipulation)

Role overview Contribute to the foundational layer for robotics AI by helping build a simulation platform that functions like a game engine designed for robots rather than players. The work focuses on producing high-fidelity synthetic data at scale so models can learn robust manipulation and perception skills, then transfer those skills to physical robots with minimal loss.

Responsibilities - Develop systems that generate large-scale synthetic training data for manipulation and perception tasks. - Improve the simulation platform so every pixel, object, and interaction is controllable for training purposes. - Support the workflow that lets developers train models and push them to real robots without rewriting tooling. - Collaborate closely with a small team on ambitious problems with high ownership and thoughtful risk taking. - Build tools that genuinely save time and energy for downstream developers.

Requirements - Strong background in machine learning, with applied experience in perception and manipulation. - Comfort working on a small team that prizes ownership, focused effort, and deep thinking. - Ability to push through hard technical problems with sustained learning. - Pragmatic engineering skills to turn research ideas into reliable product behavior. - Interest in bridging the gap between simulated training and real-world robotic deployment.

Nice to have - Experience with synthetic data generation, domain randomization, or sim-to-real transfer. - Familiarity with game-engine concepts applied outside traditional games.