Engineering Manager | Applied AI

• You've run an engineering team and made it better. Direct reports, and specific examples of people who levelled up under you.

• You get results through people who don't report to you: product managers, domain experts, other teams.

• You've led at pace, on high-performing teams with real deadlines, and you know what to cut.

• At least 3 years at a company where generative or agentic AI is central to the product, not bolted onto it.

• Teams you've led have shipped agentic systems to production. Prototypes and demos don't count.

• You can hold your own on the detail: retrieval quality, evaluation, observability, agent design. You recognise an over-engineered solution when you see one.

• Located within CET +/-2.

• Nice to have: a formal background in AI or ML, experience taking a product from non-AI-native to AI-native, familiarity with LangGraph, MCP, LangSmith or Langfuse, or legal tech and other regulated domains.

We're looking for an Engineering Manager to own AI at one of the fastest-growing AI legal tech companies in Europe. Over 2,000 lawyers use JUPUS every day, and the AI behind it is already in production and already in front of customers. You'd own the team, the technical roadmap, and the standard everything gets built to.

You'll report to our Head of Engineering and lead our AI team. This is a leadership role, not a senior engineering seat with a lead title. You won't be shipping features yourself, and that's deliberate- we want someone whose leverage comes from the team they build and the calls they make.

Legal work is document-heavy, judgement-heavy, and almost untouched by AI. We've trained on over a million German court rulings and handled over a million cases, so the hard part isn't proving AI belongs here. The hard part is doing it well enough that a lawyer will put their professional judgement behind the output.

Your job is to make that repeatable. The field throws up something new every week and most of it is noise. We need someone who can tell the difference, turn that judgement into a roadmap, and build a team that can actually deliver it.

• Make the team better than you found it: hire, mentor, and raise the bar on what good looks like here. Your impact shows up in what they ship when you're not in the room.

• Set the engineering standard for AI work: evaluation, testing and observability, and a delivery rhythm that holds up when the pressure is on.

• Own delivery at startup pace, without trading away the accuracy lawyers depend on.

• Own the technical roadmap: architecture, agent frameworks, retrieval, evaluation and observability. You won't be writing the code, but you will be the person who says whether the design is right.

• Partner with product on what gets built: product owns the product roadmap, you own what is technically possible and what it will actually cost.

• Separate signal from noise: track how generative and agentic AI is actually evolving, and turn it into decisions rather than slide decks.

• Get AI into real product workflows, working with design and engineering so it lands as something people use rather than something we demo.

• Make the rest of the business better at AI. You're the authority here, and the goal is that other teams get sharper because you're in the building.

• Explain AI to people who don't build it, from engineers to the CEO, without dumbing it down or hiding behind it.