Full-Stack Product Engineer
Summary
Builds and iterates the startup’s AI product for small-business owners, analyzing usage data to guide feature priorities and growth bets.
£70k – £85k • 0.1% – 1.0%
London, Hybrid role, contract to perm
We are partnering with an ambitious seed-stage AI startup based in London to hire their first dedicated growth hire — a builder who ships the product, reads the data, and calls the next bet.
What they doThey're building the senior AI employee for small-business owners. It plugs into the tools they already use, watches what they commit to, and drafts the work they owe — client updates, proposals, reports — from their real numbers, ready to approve. No prompting, no setup.
Small, senior team. Shipping weekly. Almost everything is still to be built, and you'd be building it.
The roleYou're a full-stack engineer who is equally at home in the data. You build the feature, instrument it, query the numbers yourself, and come back with a verdict — not a dashboard for someone else to interpret.
Roughly 70% building, 30% analysis. No PM hands you a spec. You find the highest-impact bet and go.
This is a founding-level seat: you'll shape how the team builds, measures and decides, and you'll grow the function around you as the company scales.
What you'll build and own- Product, end-to-end. TypeScript, React, APIs (GraphQL or REST), and the database underneath. Onboarding flows, connector fixes, CTAs, empty states — shipped in days, not sprints.
- The numbers. A clean event taxonomy in Amplitude (or similar), SQL you write yourself, and analysis the whole team trusts. A gap in the data is yours to fix, not someone else's to flag.
- Experiments that actually settle things. A/B tests behind feature flags (GrowthBook, PostHog, LaunchDarkly, Statsig) plus the statistical read — significance, sizing, guardrails.
- The activation funnel. Find where users lose momentum, form the hypothesis, ship the fix, prove it worked.
- Retention and growth loops. Lifecycle messaging and re-entry flows (Customer.io, Braze, Iterable, or your own), and the invite/seat/sharing loops that turn one user into their whole team.
- A builder first. Production TypeScript and React, comfortable across the stack, writing clean, reviewable code at pace.
- Genuinely analytical. You design the tracking plan, not just query it. SQL is a daily tool, not something you ask someone else to handle.
- Someone who's led. You've managed or mentored engineers, or owned a product area with people around you — and you're ready to do that again as this team grows.
- Founder instincts. You care about the business number, not the feature. Bonus if you've shipped something of your own.
- AI-native. You build with agentic coding tools daily — Claude Code, Cursor or similar — with your own workflows.
- Close to the customer. You want to talk to real users, see where they dropped off, and build the fix yourself.
Nice to have: design taste for the surfaces you ship · chat-native products (Slack, Teams) · OAuth integrations and connectors · LLM and agent products, prompt engineering, evals · Datadog or similar.
The details- Full-time, permanent, flexible hours. Two days a week together in the London office.
- You'll need the right to work in the UK.
- Four stages: intro call → technical and product conversation → practical working session → founder conversation.
Send your CV plus anything that shows how you think and build — GitHub, portfolio, products you've shipped, or a short note.
Excited by the role but don't match every line? Get in touch anyway.