Founding Product Engineer

We are a small team building Cook.ai: an AI marketing workspace that produces content, runs ad campaigns, builds funnels and works customer records for one-person businesses and agencies.

This is a founding engineering role. It is broad on purpose.

What you would work on

• Whole features, end to end. The interface, the backend, the agent behaviour, the launch and the numbers afterwards. No handoffs to a separate team, and no tickets written for you.
• Agent systems that touch real money. Campaigns with real spend, customer records with real people in them. The correctness bar is higher than a demo-grade agent and the failure modes are more interesting.
• The product surface. A workspace people sit in all day. Interface quality is not decoration here; it is most of whether the product is trusted.

What we look for

• You have shipped software that other people depended on, and you can point at it.
• You are comfortable across the stack — a typed frontend, a real backend, and the operational side of keeping both up.
• You have built with language models beyond a demo, and you have opinions about where they fail.
• You verify your own work by opening the running product, not by reading the diff again.
• You write clearly. Most decisions here happen in writing.

What we do not look for

• A specific number of years
• A computer science degree
• Experience at a company whose name we would recognise

We will ask you to talk us through something you built and why it is shaped the way it is. That conversation tells us more than a résumé does.

How we work

Remote and asynchronous, with few standing meetings. Decisions are written down so anyone can catch up without a call.

Work is judged on what it does in the running product. A change is finished when it is live and there is evidence it works — not when the code looks right. We build heavily with AI agents ourselves, and we are candid inside the team about where that helps and where it produces confident nonsense.

Compensation

Competitive salary plus meaningful equity. We will discuss specific numbers in the first conversation rather than after four rounds, because wasting your time on a process that ends at a number mismatch is rude.

Interview process

1. A 30-minute conversation about what you have built
2. A paid work sample on something close to the real product, sized to a day
3. A conversation about the work sample, and a candid discussion of the trade-offs

Three steps. We aim to complete it in under two weeks.