Member of Technical Staff, AI & Robotics

Artificial intelligence is still largely confined to the screen. We intend to bring it into the physical world.

Darwin is building humanoid robots for real-world work. We believe the key to embodied intelligence is deployment: putting robots into the world, learning from real experience, and continuously improving.

First deployments begin in early 2027.

Our goal is embodied AGI for the benefit of humanity.

Darwin is building humanoid robots for real-world work.

We believe the key to embodied artificial general intelligence is deployment: putting robots into the world, learning from real experience, and continuously improving.

As a Member of Technical Staff, you will help build our robot end-to-end, from the AI that controls them to the hardware and systems that bring them into the physical world.

• Robot Learning: Adapt vision-language-action models for robot control and train action policies using imitation learning, diffusion, flow matching, and related methods.
• Robot Systems: Integrate sensors, actuators, compute, electronics, and software into reliable physical robots.
• Mechanical Engineering & Mechatronics: Design and prototype mechanisms, structures, joints, actuators, and electromechanical systems.
• Perception & Control: Build perception, planning, and control systems for reliable interaction with the physical world.
• Real-World Data: Build systems for collecting and learning from large-scale robot experience.
• Simulation & Evaluation: Build environments and evaluations for measuring and improving robot capabilities.

We seek exceptional, high-agency engineers and researchers who can hit the ground running.

You have deep expertise in at least one area relevant to building intelligent robots, including robot learning, mechanical or electrical engineering, perception and control, or robotics and ML systems.

You move quickly from ideas to working systems and are comfortable testing your work on real hardware.

If you believe you can make an exceptional contribution to the mission, we want to hear from you.

Candidates must be authorized to work in the United States. We do not currently sponsor work visas.

On the intelligence side, we work on vision-language-action models, imitation and reinforcement learning, diffusion and flow-matching policies, large-scale robot data infrastructure, simulation, evaluation, and distributed training. On the robot, we work across perception, controls, motion, embedded systems, electronics, mechatronics, and mechanical design.

Our stack includes Python, PyTorch, C/C++, CUDA, ROS, simulation environments, embedded systems, and the systems required to turn real-world robot experience into continuously improving policies.

1. Introductory Call (1 hr)
2. Technical + systems design challenge (1hr 30)
3. Paid one-week trial, in person