Lead Product Engineer

We're an office-first team based in central Bristol. Our office is a short walk from the harbourside, green space, and several food markets.

The whole engineering team sits together, so getting unblocked is quick and sharing ideas is easy. We favour instant, on the spot collaboration over Slack messages and recurring meetings.

We aim for sustainable intensity without burnout. On rare occasions like a production incident or a critical launch, we may need to work outside of normal hours, but this is the exception. When we breach a sustainable pace, we treat the cause as a problem to fix.

You manage your own time, too. Errands, appointments and life admin during the day don't need approval or apology, as long as the work gets done.

AI is not a side project here, it is how we work. We've built an organisation of AI agents that runs alongside the human team. They sync our data, handle the tedious updates that used to eat whole afternoons, surface what actually needs a human, and even coach us with feedback on our own work. The result is that you spend your time on the things that genuinely need you, not on admin a machine should be doing.

You'll be expected to use these tools every day, to lean on them hard, push their limits, and help us find the next thing they can take off our plates. We're genuinely at the edge of AI tooling adoption and we move fast. If that excites you, you'll love it here. If the idea of working this closely with AI makes you uncomfortable, this honestly isn't the place for you, and that's okay.

In engineering, agent-driven development is the default. We delegate real, multi-step work to coding agents, manage their context carefully, and often run several at once. When an agent goes off piste, we don't just step in, we ask how can we give it the context to get it right next time. Hand-writing code line by line is increasingly the exception rather than the rule, and the nature of the job has shifted from doing the work to directing and verifying it.

That raises the bar on judgement rather than lowering it. You're accountable for everything that ships under your name, including the parts an agent produced that you didn't type yourself. The rule is simple and applies to everyone: you never merge what you can't defend. If you don't yet understand it, you build that understanding, add verification you trust, or escalate, before it goes out. We've found that this way of working produces not just better software, but better engineers as we learn on the job more than before.

We're a small team, so you'll have real ownership of your work and the freedom and responsibility to make the key decisions yourself.