Senior Backend Developer (Python, Java)

Legal AI is only as good as the knowledge behind it. At Lefebvre Sdu, we've been building that knowledge for over 200 years. Day in, day out. Law by law. Word by word. As the Dutch heart of Lefebvre Group, the European leader in legal and tax knowledge, we're also developing our own platform: GenIA-L, where lawyers and tax advisors don't just get an answer, but an answer they're willing to attach their name to. More than 75,000 professionals use it daily, generating over 100 million euros in ARR across Europe. We're looking for a Senior Backend Developer to help build that.

HOW YOU'LL MOVE US FORWARD

Our Knowledge Team is the layer between the content and the AI. We pull in millions of documents from internal databases, public sources, and the web, standardize them, and make them usable for the AI agents that distill answers from them. The real challenge isn't storage but selection. Out of those millions of documents, exactly one paragraph needs to surface the moment a lawyer asks a question. Surface the wrong one, and the agent gives an answer that sounds plausible but is wrong. In our field, that's not a bug: it's a liability issue. What you build determines whether the answer holds up.

There's work to be done, and it's the kind of work you could talk about for a whole evening. The pipeline has been running for six years, built in Scala and Java, and parts of it deserve a new foundation. Reranking is the biggest need: determining which passage, out of a mountain of legal documents, actually answers the question being asked. Not just which one contains the matching words. That's the problem we’re solving. The team is strong on the scientific side, in retrieval, and in understanding what a model can and can't handle. What we're looking to add is engineering firepower: someone who can turn ideas into systems that deliver. Not to rebuild everything over six months just because it was once set up differently, but to see where the real gains are and start there.

• You build and improve the pipeline that turns content into knowledge our AI agents can use.
• You work on reranking quality, so the agent gets the right information, not just the most information.
• You help shape the approach to retrieval and knowledge extraction, bringing in what you've seen work elsewhere.
• You work in Python for everything new, and in Java where the existing codebase calls for it.
• You make technical decisions, even when no one is looking over your shoulder.