Senior Reinforcement Learning Engineer

About the role

The autonomy team at Gravis Robotics builds autonomous systems for excavators operating in real construction environments. You will develop data-driven planning and control systems that generalize across machine models and soil conditions, while addressing the challenges of sim2real transfer and deploying robotic systems in production. This role involves integrating learned components into a larger software stack and collaborating with cross-functional engineering teams.

Responsibilities

• Develop data-driven planning and control systems for autonomous excavation.
• Contribute to simulation improvements to address the sim2real gap.
• Define data collection and curation pipelines for policy training.
• Design experiments to improve performance and robustness.
• Explore adaptive and online reinforcement learning in deployed systems.
• Provide mentorship and supervision for junior team members and interns.
• Integrate learned components into the larger software stack.
• Build tools for analyzing and evaluating the behavior of learned components.

Required skills

• Reinforcement learning
• Python
• PyTorch
• C++
• Robotics
• Sim2real
• GPU accelerated simulation

Nice to have

• Hydraulic machinery
• Supervised learning
• Imitation learning
• ROS
• Data curation
• Mentoring

Qualifications

• 2–5 years industry experience developing Reinforcement learning systems for control and/or planning and deploying them on real robots

About the Company

Gravis Robotics is a high-growth Series A start-up backed by SoftBank, bringing Physical AI to the construction industry by turning heavy construction machines into autonomous robots. Originally an ETH Zurich spin-out, the company develops machine-agnostic retrofit kits that add autonomy to excavators and wheel loaders.