Careers
Build the Context & Control Model for AI operations. Fully remote, with a senior team that ships, owns, and stays close to the customer.
Why this work
A quiet bet on production.
AI agents are getting good at writing code. They're nowhere near safe enough to operate the systems that run on it.
The gap is the Context & Control Model. Context: a typed, time-versioned model of every change, dependency, and incident in production, so an agent reasons instead of guessing. Control: every action an agent proposes is checked out-of-band, inside a hardware-isolated boundary, before it ever touches production. Reasoning without a boundary is not safety. We're building both halves, end-to-end.
If you've spent your career on the parts of software that don't make it onto landing pages (schemas, ingestion, query planners, change spines, execution boundaries, audit trails) and you want to build the foundation the next decade of operations gets built on, we'd like to meet you.
How we think and work
Six values. Held seriously.
How we work is part of the product. Async-default, in the field, low-coordination. These are the values we hire and ship by.
Open roles
Don't see your role? Send us a note →
Come build this with us.
If it's not the right title, send us a note anyway.