Research Engineer - Robotics

• Work across simulation, learning, control, software and hardware to close the loop between research and real robots.
• Build and improve the stack we use to train, evaluate and deploy robot policies and controllers, in simulation and on hardware.
• Work with research teams to apply new ideas to working prototypes on real robots.
• Make advanced robotics pipelines robust on real robots.
• Set priorities and make sound technical decisions when the problem is ambiguous, and solve real problems alongside researchers.
• Build the software that connects models, agents and robots: libraries, interfaces and tooling.
• Deploy your work on real robots in experiments inside and outside the lab.

• BS, MS or PhD in robotics, computer science, engineering, mathematics, physics or equivalent practical experience.
• Preferably >2 years of experience.
• System-level thinking: you reason about how software, hardware and algorithms interact, and you can trace a failure across layers.
• Either a research background with deep technical expertise that you enjoy applying to real robots, or experience in applied research: working with R&D teams and using state-of-the-art methods in physical AI.
• A solid understanding of robotics theory and math: kinematics, dynamics, control, estimation and optimization.
• Hands-on robotics experience: you build, debug and run things yourself, in code and on robots.
• Real hardware experience is highly preferred.
• Strong C++ and Python.
• Strong Linux skills.
• A collaborative mindset, working closely with researchers, engineers and operators.
• Contributions to open-source robotics software.

Note: Candidates are not expected to meet every qualification in the above list. If you do not have all of these qualifications you are still encouraged to apply.

Our stack includes the following components. We look for T-shaped people: broad across the stack, with deep expertise in one or more of these areas.

• Reinforcement learning for robotic control. Reinforcement learning in simulation with zero shot sim2real deployment.
• Agentic pipelines. High level agentic reasoning and planning pipelines for long horizon tasks.
• Perception and Spatial AI. Perception and spatial understanding for robots: calibration, state estimation, and 3D scene representations that feed policies and agents.
• Robot software. The interfaces between hardware and software, and the tooling used to operate, debug and test robots to accelerate research.

Our Mission

Our mission is to solve the most important and fundamental challenges in AI and Robotics to enable future generations of intelligent machines that will help us all live better lives.