Machine Learning Founding Engineer
My client is hiring a Machine Learning Founding Engineer to help build humanoids that do real, useful work, not just polished stage demos. Their mission is to get robots out of the lab and into the real world.
If you’ve shipped robots before, thrive in fast, hands-on environments, and care about RL, sim-to-real transfer, control latency, and safety, we’d love to talk to you.
About the Role
This is a hands-on role where you’ll work at the intersection of control, hardware, simulation, and real-time systems. You’ll spend most of your time with real hardware, running experiments, collecting data, and iterating quickly. Simulation will be a tool, not the center of gravity.
As a founding engineer you will be a critical part to the company direction and growth.
What You’ll Do
- Deploy humanoids into real environments to perform meaningful, sustained work.
- Ship to real hardware almost daily — experiment, debug, and iterate in the loop.
- Use simulation strategically to accelerate learning, but lean on real-world testing.
- Build strong data and evaluation loops using teleoperation, VR rigs, and other signal sources.
- Develop and maintain robust control stacks that are fast, stable, and safe.
- Reproduce and ablate new research quickly, bringing effective methods into production.
- Work closely across software, hardware, and controls teams.
Tech Stack
- NVIDIA IsaacSim / Isaac Lab
- PyTorch
- Weights & Biases
- Python, WebXR, Vuer
- Linux + GPUs
What I'm Looking For
Must-Haves
- Proven experience shipping real robots (not just simulations).
- Strong background in reinforcement learning or other high-stakes control problems.
- Strong Python skills and comfort in Linux / GPU environments.
- Hands-on, problem-solving mindset.
Nice-to-Haves
- Whole-body or model-based control, VLAs.
- Experience building safety layers in control stacks.
- Teleoperation, VR interfaces, or human-in-the-loop systems.
- Real-time systems, sensor fusion, trajectory optimization.
Why Join
- Work on real, high-impact problems in humanoid robotics.
- Ship to hardware fast and often — minimal bureaucracy, high velocity.
- Collaborate with a small, senior team of builders and researchers.
- Own systems end-to-end: from teleop rigs and control stacks to deployment.
- Your work won’t end up in a demo reel — it will make robots genuinely useful.
Compensation
$150k - $250k DOE + significant equity stake
Location
This is a full-time, on-site role in San Francisco. Relocation support is available for strong candidates. Remote candidates may be considered but must be willing to travel to site frequently.
