Own the outcome
Small teams, real responsibility, and a short line from your work to a customer's action in production.
Careers
Build the step that comes after the prediction.
Plenty of teams can produce a score. Almost none have made it safe for software to act on one inside a regulated enterprise. That is the problem we work on, and we are hiring people who want to own it.
What it is like here.
Honesty over hype
We report reproducible numbers and label synthetic ones. The same standard applies inside the building.
Production is the craft
The interesting work is the unglamorous part: governance, safety, isolation, and the action itself.
Close to customers
Engineers and researchers sit with the data and the deployment, not three handoffs away from it.
Teams we are growing.
Graph ML & research
Relational foundation models, temporal safety, and the science behind predictions agents can act on.
Platform & infrastructure
Single-tenant deployments, on-prem and air-gapped, the gateway, and the audit and isolation layers.
Applied delivery
Forward-deployed work: standing up use cases on customer data and closing the loop in their environment.
Security & compliance
Threat modeling, fail-closed design, and the controls that make autonomy defensible.
Product & design
Make agentic action legible: the policy, the human-in-loop review, and the audit trail.
Go-to-market
Help regulated enterprises understand what changes when the score becomes an action.
No role that fits? If you are excellent and the mission resonates, tell us how you would help.
Introduce yourself.
Send us who you are and what you would want to own. We read every note and reply if there is a fit.
We read every note and reply if there is a fit.