Senior Reinforcement Learning Engineer

This full-time position in Zurich involves developing data-driven planning and control systems for autonomous excavation tasks. The role requires building control modules that function across various machine models and soil conditions on real construction sites. You will also work on improving simulations to mitigate the gap between simulated and real-world performance. The job demands deep expertise in Python along with a good level of proficiency in C++. Candidates should possess a background in data-driven planning or control within robotics. Experience working with physical robots is essential, specifically regarding the challenges of transferring skills from simulation to reality. The employer seeks individuals who have deployed robotic systems in production environments. Gravis Robotics is a Series A start-up backed by SoftBank, originating from ETH Zurich, that uses learning-based automation to turn heavy construction machines into autonomous robots. Their technology allows one operator to manage a fleet of earthmoving machines through a gamified interface. The company’s retrofit kit adds autonomy to excavators and wheel loaders using LiDAR, cameras, and edge compute. This role suits a roboticist interested in applying reinforcement learning to physical AI within the construction industry.