Full Stack Engineer
About us
We build AI that runs a restaurant’s delivery business.
Most restaurants now make a serious part of their money through delivery platforms, and almost nobody is really managing that side of things. Menus drift out of date, stores sit closed while the kitchen is open, and problems surface a week later in a report, if they surface at all. Our product watches all of it, works out what is costing the restaurant money, and tells the owner what to do about it. More and more, it just does it.
Real customers run their business on this every day. It keeps us honest: when we get something wrong, somebody feels it the same evening.
Small team, everyone owns what they build. Remote first. Our own product, built for the long term, no client work.
What you’ll do
- Build product features end to end: React on the front, Node on the back, all the way into production.
- Turn vague asks into decisions. Clarify the problem, cut the scope, pick the fix.
- Write the tech design for what you build.
- Work with delivery platform integrations and their APIs.
- Deploy and run your own services on GCP.
Direct agents. Claude Code, Codex and Cursor do the typing. You do the thinking, the review and the verification.
What we expect
- 3+ years of commercial fullstack work: React, Node.js and PostgreSQL in production.
- You already work AI first. This is not something we will train you into. We need a year or more of daily, hands on work with Claude Code, Codex, Cursor or similar. You set the context, you review every line, you own the result.
- You have shipped things real users depend on, and you were around when they broke.
- Production experience in the cloud, GCP preferred, and enough DevOps to deploy and debug what you build.
- You can get from a business problem to a product decision instead of implementing the ticket literally.
English B2+ and clear writing. Most of our decisions happen in text.
What matters more than years on a CV
- You finish things. Given a task with rough edges, you close the gaps and deliver, instead of handing it back with questions attached.
- You ask early and precisely. A question at hour two costs five minutes. The same question on day four costs a day.
- You know when to push back and when to commit. If something is genuinely wrong, say so clearly, once. Then we decide and you run with it.
You automate your own work. If you have done something by hand more than twice, we expect you to automate it rather than get faster at it.
Nice to have
- GraphQL in production
- Terraform or another infrastructure as code tool
- Food tech, delivery or POS
- Your own agents, MCP servers or dev automations
React Native
How we work
Small team, very little process, short loops. Two week sprints. Agentic engineering is the standard here rather than an experiment, which makes sense given what we build.
Engineers sit close to the product. You will talk to support sometimes, and you take your own features through QA and release yourself. Reliability matters more than speed on anything that touches orders, and you own what you ship.
You will get through more work here in a quarter than in a year at most places. Nothing waits on a committee, and the agents handle the typing, so the hard part of the day is judgment rather than boilerplate.
Remote first.
Good fit if you want real ownership, fast feedback, and visible impact on something people use to run their business every day.
Hard fit if you want a narrow scope, a calm pace, or a lot of process between you and production.