Build evaluation tools for physical AI.

We work across robot labs, station hardware and evaluation software. If you want to help build Prism, show us your work and tell us where you can contribute.

What we work on

Prism connects physical scene preparation to the software used to run, score and compare robot evaluations. The work spans reliable station operation, perception, evidence capture, analysis and clear interfaces for lab teams.

We value people who can investigate a problem, build a useful solution and follow it through to the person using it.

Areas of contribution

Robotics & perception
Scene verification, object tracking, policy evaluation and the connection between observations and physical conditions.
Software & systems
Station services, evaluation workflows, test libraries, integrations and data that stays understandable across runs.
Hardware & optics
Projectors, cameras, mounting and the practical requirements of dependable workcell hardware.
Product & design
Interfaces for operators and researchers, evidence review, comparison tools and product communication.
Lab operations
Task setup, repeatable procedures, evaluation quality and documentation that helps the next operator.

Share something you have built.

Send a project, repository, paper, physical build or concise write-up. Explain the problem, your contribution and what you learned. Relevant work can come from research, employment, independent projects or other paths.

Get in touch

Email hi@lightcompany.ai with a short introduction, relevant links, your location and the kind of work you want to do.

We welcome expressions of interest. Specific openings, location and employment terms will be discussed directly. Please do not send identity documents or sensitive personal information in an initial message.