Picking and Packing Staff Required in London

This company sits in a part of the AI ecosystem that most people haven’t really seen yet. They’re not building a single product; they’re building the layer that sits behind how AI systems are evaluated, improved, and ultimately trusted in real-world environments.

They’re already working closely with some of the most advanced AI teams globally, helping tackle a problem that’s still wide open: how do you reliably evaluate whether an AI system is actually performing well, and how do you generate the data that makes it better over time?

The deeper layer is what they’re building internally. They define what “good” looks like, design evaluation frameworks, and create feedback loops that directly improve how models behave. That’s the part that actually shapes how systems get better in production.

The work feeds into how modern AI systems are being tested, evaluated, and iterated on in practice — not just in theory. You’re seeing how different teams approach building AI systems, how evaluation and data actually drive performance, and how things move from something that looks good in a demo to something that works reliably in production.

You’d be working directly with the founders, helping define how these systems are built and evaluated, while also contributing to product and technical direction. This only really makes sense if you view it as a period of acceleration, not a long-term lifestyle.

If you’re early in your career, already building AI systems, and motivated by getting very good, very quickly, this kind of environment can compress years of learning into a much shorter window.