Wabilon is building AI-powered software to transform how technical, regulatory and urban planning documentation is reviewed. Our product helps professionals and public institutions analyse complex documents, detect inconsistencies, and support decision-making through intelligent automation. We are looking for a Software Engineer - AI Systems to work directly with our CTO, and play a key role in the development of our core product.

We are two cofounders. You will be the third person. The work You will work with large, messy technical documents and check them against a structured corpus of regulation. This involves:

Long-context LLM systems, where cost and latency depend on how much context we carry and how often we use it.

Evaluation systems, where the main failure mode is not a wrong answer, but a missing one. Recall matters.

Agentic pipelines that may run for hundreds of steps and need to be reliable, debuggable and affordable.

Retrieval, document understanding, structured reasoning and validation over complex regulation.

The product around it: APIs, interfaces, infrastructure and internal tools that make the system usable in production. What we look for No years-of-experience requirement, and we won't ask how long you've used any framework. Instead, we want to know about:

LLM systems you shipped or contributed to,



and how they failed, specifically.

How you handle evaluation, especially where being wrong has consequences.

Your familiarity with infrastructure: where cost and latency come from, what breaks under load.

Your familiarity with current techniques for agentic workflows and context retrieval, and their failure modes. We value slope over starting point. How fast you get good at something new matters more than what you already know.

What we offer A senior role with real ownership over the core product.

Barcelona-based, hybrid, with regular time together in the same room.

Direct collaboration with the founding team.

Competitive salary based on experience.

Equity upside for the right long-term person, subject to role, commitment and plan terms.

A culture of autonomy, ambition and excellence.

Employment contract preferred. Freelance is also possible if it fits better. How we hire First, a 30-minute conversation.

Then, a 60–90 minute working session on a potentially real problem.

Bring whatever AI tools you use daily. We use them too. We are interested in how you decompose the problem, where you disagree with the model, and how you check whether an answer is actually right.

📌 Senior Engineer, AI systems (Barcelona)
🏢 Wabilon
📍 Barcelona