Build the platform a production runs on: generation, storage, permissions and the interface around them. Small team, wide surface, work that ships the same week.
What you’d do
- Ship features end to end — schema, server, interface — and own them afterwards.
- Work on the parts where correctness matters: permissions, credits, versioning, money.
- Keep frontier models useful inside a real workflow, which is a different problem from calling them.
- Read what breaks in production and fix the cause, not the symptom.

Why it’s hard
The surface is wide and the users are demanding: they look at frames all day, so a misaligned label reads as a broken tool. And a bug in permissions or in credits is not a bug you get to fix quietly.
What it isn’t
Not a role with a narrow slice and a spec handed down. Not a place where the interface is someone else’s problem.
Who tends to fit
- You have shipped something you were responsible for end to end, and maintained it afterwards.
- You are strong in TypeScript and comfortable in Postgres — including the part where you have to reason about who can read a row.
- You care how it looks and feels, and you would rather cut a feature than ship it half-finished.

📌 Developer (Milano)
🏢 Greatspir
📍 Milano