Join Us and Make Embodied AI a Reality

We’re building the backbone for general-purpose robotics: a platform where imitation learning, reinforcement learning, and foundation models converge to power dexterity on real hardware. And we’re making it usable by anyone, no PhD required.

At the center of this vision is:

• A powerful cloud infrastructure for training thousands of custom visuomotor skills

• An interactive AI agent that guides users through every step: hardware setup, data design, model selection, training, online fine-tuning, deployment, and validation

 Our goal: make training robots intuitive, fast, and cost-effective.

 This is a fast-moving startup: small team, big impact. You’ll ship real features, work hands-on with robots, and help define the future of embodied AI.

Roles Open

AI Engineer (Seattle, hybrid)

Enable interactive and scalable training of visuomotor models and VLA architectures for real-world robotics. Your expertise will power our AI agent, helping it guide users through data design, model training, and deployment.

Expertise:

• Training and deploying visuomotor policies on real hardware

• Imitation learning and reinforcement learning

• Automated data curation

• Automated evaluation and validation

Cloud Engineer (Seattle, hybrid)

Build self-service infrastructure for massive datasets, distributed training, and real-time telemetry, so developers can train and deploy robot skills without friction.

Expertise:

• Scalable APIs

• Training infrastructure

• Video data pipelines

• Cloud orchestration

Robotics Engineer (Seattle, hybrid)

Design control systems for diverse hardware platforms that enable end-to-end training loops, both offline and online. Your work will make robots learn new skills safely and reliably.

Expertise:

• Inference and rollout of visuomotor policies on real-time hardware

• Robot control, including force-based control

• Computer vision and camera calibration

Curiosity and Initiative Required

No resumes. Just tell us what you’ve built, what you can contribute and why you want to join, in a 3 minute recording (audio or video).