The role
A team of generalists builds the platform that every Fig deployment runs on. You are one of them. There are lanes rather than titles: agent runtime and evals, the connector framework, per-customer deployment tooling for VPC and on-prem, SSO, RBAC and audit, and the product surface customers see. You pick a lane, own it for a while, and rotate when the work calls for it.
The feedback loop is short. Something an engineer hand-built at a customer last week is your job to turn into a platform feature the next ten customers get for free. What breaks in the field gets fixed in the product, by you, usually within the week.
What you'll do
- Build and run the systems every deployment depends on: the agent harness, the eval framework, the connector layer, and the tooling that stands a customer environment up in hours.
- Turn field-built pieces into platform features with the tests, docs and rollout plan that let a Forward Deployed Engineer use them without asking you.
- Own production: tracing, metrics, safe rollouts, incident response, and the latency and cost tradeoffs of running frontier models for real work.
- Ship weekly. The platform is in production at customers every day, and so are your changes.
- Spend time with the deployment pods, on site when it helps, so the platform is shaped by what customers actually run.
What we look for
- Four or more years of experience, strong in at least one of Python, TypeScript or Go, and able to read the other two.
- Has run production systems, not only built them. Knows what an on-call week feels like.
- Has shipped LLM features to real users and has opinions about what went wrong.
- Builds end to end. If it needs backend, infra, a front end or a prompt, you own it.
- Writes code other people can change. Clear interfaces, clear names, small pull requests.
Nice to have
- Experience with enterprise identity and permissions: SSO, RBAC, audit logging, data residency.
- Time at a company where the product was deployed into customer environments rather than hosted once.
Location
San Francisco, in the office four days a week with the platform team.
About Fig
Fig's mission is to change how businesses operate. We build AI systems that run inside companies, on their data, their rules and their choice of models, and we deploy them with a team that stays until the work is running. More than 300,000 professionals in over 70 countries already work with Fig every day.
We are majority employee-owned. There is no outside board setting the pace and no fundraising cycle to perform for. We answer to our customers and to each other, and the people who build the company own it.
The team is small and talent-dense. We like people who are direct, who would rather ship than present, and who take a customer's problem personally.