The long version

I build AI systems, and I can tell you whether anyone actually needs one. Those turn out to be different skills.

I started on a sales floor at Tingis, calling people who had not asked to hear from me. What that work teaches you is not persuasion. It is that people rarely state the thing they actually want, and that your job is to find it before you propose anything.

At devUP I moved into delivery, managing software projects across two teams. The same lesson arrived at a larger scale: the projects that fail usually fail because nobody translated the brief, not because the engineering was beyond anyone.

Since starting my master's I have been building the systems myself. AI agents, both hand-written and orchestrated through n8n. Machine learning models. Fine-tuning and retrieval-augmented generation. What pulls me is the gap between a vague ambition and something specific enough to evaluate.

I am looking for my next role. The problems I want are the complex, quantitative ones: the sort that are not yet well defined and will not survive a guess.