Full Stack Engineer - SF, Hybrid

Many of our customers work outside San Francisco. When one of them says "it would be great if the dashboard did X", we'd like the person who can build X to be in the room.

That's you.

At each::labs, our mission is to close the imagination gap. One API across 600+ image, video, audio and 3D models, used by 20,000+ developers. Under the hood are the dashboard, playground, SDKs, billing and API that make it all feel simple. Most of our team is in Istanbul. You'll be our engineering presence on the West Coast.

You'll ship features end to end, across the frontend, backend and API. You'll also be the engineer closest to customers. That means joining a sales call to answer the hard question, then shipping the fix before the follow-up email goes out.

You'll work with the wider engineering team, and with product on what we build next.

• The first five minutes. Signup → API key → first generation should feel like magic. Every step between "create account" and "my first video" is yours to question, measure and remove.
• Costs developers can trust. 600+ models, priced in different ways: per second, per image, per run. Developers should see exactly what they spent, on which model, and why, and the dashboard, the invoice and the API should always agree to the cent.
• SDKs and docs people enjoy. Typed clients that stay in sync with the API, quick starts that work the first time, and error messages that tell you what to do next.
• Agents as first-class users. More and more of our traffic comes from agents, not people typing code. That means an MCP server that exposes 600+ models cleanly, docs and specs an LLM can read and act on, tool descriptions that get the right model chosen, and errors an agent can recover from without a human.
• Fast and boring under spiky traffic. One customer's launch can multiply our traffic in minutes. Queues, retries, provider fallback, rate limits, idempotency, and p99s that don't move.

• Owned a product developers use: a public API, an SDK, a CLI, or a dashboard people live in
• Designed an API other people depend on, and lived with the consequences: versioning, deprecations, the breaking change you still regret
• Built something that AI agents use, not just people: an MCP server, a set of tools for function calling, or an API designed so a model can find the right endpoint and recover when it fails
• Shipped an AI-powered feature to real users and dealt with what that involves: evals, latency, cost per request, and failures that don't throw exceptions
• Built or fixed something that touches money, such as metering, billing, credits or usage limits
• Found out why something was slow and proved it with traces, profiles or query plans
• Shipped to production many times a week, and been the one who got paged when it broke
• Talked to the people using your product and changed what got built because of it
• Worked with image, video or audio pipelines (a plus, not a requirement)

We care about what you have built. If your tool, app or service fits, even if some of the language above doesn’t match your background, we’d like to talk. Send us a link.

• You're strong in a typed language on the frontend and at home on the backend. You choose tools for the problem and pick up new ones quickly.
• You use coding agents every day, but you've also been on the other side: you've designed what an agent sees, and you know a good tool schema is product design.
• You write things down: short specs, clear PR descriptions, honest postmortems.

• You own features from idea to production to the metric that tells us whether it worked
• We ship small and often, behind feature flags
• Everyone reads support tickets and talks to customers
• Incidents get blameless postmortems, not…