Lead AI engineer for legal tech

About Juris The law applies to everyone. The knowledge to use it has always been scarce: held by specialists, earned over years, delivered one client at a time. So the HR lead, the caseworker, the in-house team, the founder, all end up guessing at questions the law already answers. We are building the layer that ends that scarcity. AI can now take what lived only in a specialist's head and make it run: workflows that codify how the best experts handle an entire case type, execute it step by step, grounded in the actual law the whole way. That automation layer makes legal knowledge abundant: anyone with a legal question gets an answer they can act on, and an expert's judgment reaches a thousand cases instead of one. Juris AI has been in production with Danish companies and the public sector since 2024, built on a curated corpus of Danish legal sources with every claim traceable to its provision. We recently launched Juris OS, introducing our first automations alongside the platform they run on: Collections for a customer's own cases and documents, Workflows for the legal work that repeats, and Governance over how AI gets used. The first automation handles freedom-of-information requests end to end, and more case types are in build. Who we are looking for We are hiring one engineer. Seniority here is not years, it is what you can do with the tools that exist right now. The person we want shares the optimism underneath all of this: that AI is making expert knowledge abundant, that we are at the start of it, and that there is no better place to be than building at the front of it. Beyond that, three things matter most. You are a strong engineer first, fluent across a backend and the product on top, and the AI tooling makes you faster rather than standing in for the fundamentals. Beyond that, three things matter most. You have shipped this, and lived with it afterwards. LLM, retrieval and agent systems in production, plus the evals, the regressions, and the failure at 3am that turned out to be a chunking decision made four months earlier. Building the demo is the easy quarter of the work. You are current, and you stay current. You know what shipped in the last few weeks, you have a view on it, and you have already tried the ones worth trying. This tooling rewrites itself every quarter. Working from what was true when you learned it does not hold here. You are not attached to how you have always done it. We ship in small pods on short cycles, with AI carrying the volume and your judgment applied where it actually counts. If your instinct is a long planning phase and a three-week branch, we will frustrate each other. How we work AI is not something we ship, it is how the company runs. It writes with us, reviews with us, and absorbs the volume that would otherwise take a team several times our size. Staying small is the strategy, not a constraint we are waiting to grow out of. Grounded over plausible. In law, an answer you cannot trace is worse than no answer. We ship fast, and we stand behind what we ship. Small team, real ownership, no passengers. What you will own The AI build end to end: retrieval over Danish legal sources, agent workflows, the Python and FastAPI backend, and the product on top. Spec-first, agent-driven development. Claude Code, Cursor and agents are your default, not a side experiment. Shipping fast, then owning the verification: security, grounding, traceability. Sitting with the professionals who use Juris, making the commercial call yourself, and building what moves the business over what is merely clever. Also true of you You hold the line on security and code quality, and treat them as what lets you ship fast with AI rather than a brake on it. When agents generate code quickly, your judgment and review are what turn that into shipped product instead of a review backlog. You read the business, not just the ticket, and can judge whether a thing is worth building at all. People want to build with you again. You explain hard things simply and work well with non-engineers. You want the ownership and scope you cannot get as engineer number fifty somewhere else. Nice to have A pull toward legal, regulated or otherwise hard domains. We are based in Copenhagen and strongly prefer you here or in the area. Early on, this work goes faster in the same room, so Copenhagen is the default rather than the exception. For the right person elsewhere in Denmark or Europe we will find a way to make it work, but proximity is a real plus. Danish is a plus, not a requirement. Compensation There is a base salary from day one. It is modest, sized for where the company is now, and we expect to raise it after the coming funding round. Equity is a real part of this seat: a strong option, not a hard promise. We will talk structure and size with you before the round closes. If cash today is the deciding factor, this is not the right seat, and better to know that here than three conversations in. How to apply Send your CV, your GitHub, and the one thing you are proudest of shipping with AI tooling, plus how fast you shipped it. We read that last part first.