For the jobs where
mistakes cost more
than money.
Vertex builds foundation models for extreme-purpose robotics. Our customers are the firefighters, combat medics, hazmat teams, and search-and-rescue crews working in environments where every additional second of human presence is a tradeoff against death.
Firefighters enter collapsing structures. Combat medics treat casualties under fire. Search-and-rescue crews work avalanches, floods, and rubble. These are the environments where every additional second of human presence is a tradeoff against death — and they are the environments with the least automation.
Not because the work is unimportant. Because the work is hard. No robot has been general enough to handle the chaos of an actual scene. We are building the model layer that lets robots stand alongside the operator and take the part of the job that costs the most.
One backbone, many platforms. We train on large corpora of sensorimotor data — teleoperation, motion capture, simulation rollouts — and apply the recipes that built foundation models in language and vision to produce policies that generalize across embodiments and across the messy edge of the real world.
The data problem in robotics is solvable. The combination of high-quality human demonstration, scalable simulation, and cross-embodiment transfer is sufficient if you build for it from the start. We are.
We operate from first principles. Every design decision — network architecture, training infrastructure, hardware interface — is made by the team doing the research. We do not copy choices because they worked somewhere else.
Vertex was founded in 2026. We are a small, focused, fully remote team across the United States. We keep a tight synchronous cadence Monday through Thursday. Friday is a think day. Saturday is off, always. Read how we work.
We are hiring researchers, robotics engineers, and operators who have already done hard things and want to do harder ones. If that’s you, we want to hear from you.