Estate agency runs on conversations: the first message from an owner thinking about selling, the follow-up a week later, the reminder nobody sent. You would build the agents that handle them, in a small team at WAT in Brussels.
About Cultiv
We started with an AI assistant that generated seller leads on estate agency websites. It worked, but agencies kept telling us the hard part starts after the lead arrives: follow-up, valuations, reports, the compromis. So we built the rest of it. Today more than 100 agencies in Belgium and France work with us, and the CRM is open from the moment they arrive until they close.
The role
You own the AI layer: the widget conversation, the valuation pipeline behind it, and the follow-up that runs without anyone pressing a button. It runs on Claude and Azure OpenAI today, with automations in n8n, and you decide where it goes next. Expect to spend as much time on evaluation, and on what happens when a model is confidently wrong, as on the prompts themselves.
What you’ll do
- Run the widget conversation. It has to qualify a seller in a dozen messages, in Dutch, French or English, without feeling like a form.
- Own the valuation pipeline. Three independent methods (comparable sales, land plus construction, rental yield) each produce a number, and something has to reconcile them into one estimate with a range an agent can defend to a client.
- Write the tools the agents call: cadastre lookups, comparable sales, EPC and permit documents, the CRM itself.
- Build the evaluation harness, so that whether a prompt change helped is measured instead of read from transcripts.
- Watch cost and latency. A homeowner abandons a chat that thinks for eight seconds.
Your profile
📈Experience & background
- At least a year of software engineering, most of it on AI products.
- You have shipped an LLM system to real users and stayed responsible for it afterwards.
- A degree in computer science or engineering, or projects that show the same thing.
- Dutch native speaker.
💻Technical skills
- Python for the AI services, and enough TypeScript to work inside a Next.js codebase.
- Hands-on with the Claude, OpenAI or Gemini APIs: tool use, function calling, structured outputs.
- Experience with multi-agent orchestration, and a sense of when a single well-built prompt is enough.
- You know how to evaluate an LLM system beyond reading transcripts.
- RAG where it is useful. Much of what we handle is structured data rather than documents.
🧠Mindset
- You put something unfinished in front of a user in order to learn from it.
- You take responsibility for what you build once it is live.
- You can tell a model that demos well from one an agency relies on daily.
- You follow model releases and can judge which ones matter for us.
- You want to understand how estate agents actually work, and you ask them.
What we offer
- A competitive salary package.
- A small team. What you ship is in front of agencies within days.
- A desk at WAT, the best incubator in Brussels, alongside other early-stage tech companies.
- Full ownership of the AI layer. There is no ML team above you.
- Real data problems: cadastre records, EPC certificates, listings that contradict each other.
Our process
- A first interview online with Cultiv’s CTO (30 min).
- The technical interview at our office.
- You meet the rest of the team over lunch.