Intuition
←
intuition

Robotics Research Engineer

About the Role

You will build the systems that let our research run on real robots in the real world. Evals at scale, deployment infrastructure, the bridge between model and machine. This role consolidates model training, model inference, deployment engineering, and on-site deployment management — because at our stage, the same person ships the model and gets it onto the robot. Whatever the research team needs to test a hypothesis on hardware, you build it.

What You'll Do

  • Train and fine-tune our VLA (Vision-Language-Action) models. Reproduce, ablate, and improve across our models and the public ones.
  • Optimize and deploy models on edge devices for real-time inference. Latency, throughput, reliability, and quantization where it matters.
  • Build and scale our evaluation infrastructure. Run benchmarks in sim and on real robots, across the standard ones from industry and research.
  • Design and ship new evals where the existing ones do not measure what we care about.
  • Build the deployment stack for real robot fleets. Remote teleoperation, on-robot model distillation, cloud-to-edge inference, intervention detection.
  • Integrate, calibrate, and tune real robot platforms for production deployments. Travel to customer factories. Own the deployment from arrival to handoff.
  • Work side by side with the research team to take ideas from notebook to robot fast.
  • Use AI coding tools as a core part of your workflow.

What We're Looking For

  • Strong engineering background with hands-on robotics or embodied AI experience. You have shipped code that ran on real hardware.
  • Production experience with ML infrastructure. Training, eval, deployment, observability.
  • Comfortable across the stack. Python, C++, distributed systems, GPU.
  • Willing to travel to customer sites and work on the factory floor when needed.
  • High agency — you don't wait to be told what to do.
  • Fluent in English.

Why Join Us

You ship the model and get it onto the robot. Same person, same week. The bridge between model and machine is yours.