Robotics AI / Learning Systems Engineer

Los Angeles, CA // Full Time or Intern // On Site

We're looking for a learning systems engineer who can close the loop between perception and action and make a robot adapt in real time. This is closed-loop work, since what you build runs on real hardware under real constraints, not just in simulation.

We care about first principles over specific methods. You should be comfortable reasoning about state (often partial/noisy), dynamics, and feedback, and using whatever tools are appropriate to produce systems that are stable, responsive, and robust under noise, latency, and changing conditions.

What You Will Own

• Build closed-loop systems mapping sensor inputs to actions or control updates.
• Design around state estimation, dynamics, and feedback (not just reactive policies).
• Connect perception outputs to motion or process control in real time.
• Develop and tune feedback loops for stability, responsiveness, and robustness.
• Reason about system behavior (latency, noise, drift, distribution shift) and adapt accordingly.
• Iterate quickly on physical systems.

Evidence We Want To See

• Show closed-loop systems where perception drives real-world action.
• Work grounded in system fundamentals (state, dynamics, feedback).
• Show systems that ran on real hardware, not just simulation.
• Show measurable improvements to stability, performance, or adaptability.
• Show evidence of diagnosing and fixing real-world failure modes.
• Show ability to handle partial observability, noise, or distribution shift in practice.