Learn how things work. Build what's missing.
Most problems start as data nobody has really looked at. Study it until it makes sense, then build something that works.
Three habits, every problem.
It doesn't matter much if the problem is a product, an experiment or a warehouse. I go through the same three steps, and I go through them again.
Before deciding anything, I want to see how the thing really works: the process, the people, the numbers. Most problems get clearer after the right question than after the first solution.
Most of what's needed already exists somewhere: a method, a library, a tool, someone who solved it before. The work is choosing the right pieces and making them fit together.
A before and an after, agreed at the start. Then small steps, each useful on its own. When something works, the loop starts again on the next question.
AI changed how fast I can move, not what I'm responsible for. It lets me research, compare and build on several fronts at once. Deciding what is worth building, and checking it against reality, stays with me.
Same approach, different problems.
Not a CV. Four places where I met a new kind of problem, and what the three habits looked like there.
A bachelor's and a master's in product innovation. Where I learned that a problem is something to design, not just to solve, and that the first prototype is a question, not an answer.
A year of PhD research at DTU, inside a large European project on industrial 3D prototyping and simulation software. One publication, and a habit I kept: no idea is true until the data says so.
With my family I co-founded Athesis, a B2B import company. I took on technology and operations and, alone and with AI, built the platform the whole company works on, three times, each time when the previous one reached its limit.
I'm curious about what this approach does in places I don't know yet: other companies, other teams, other countries. That's the next chapter, and it isn't written.
How I evolve, versioned.
I treat myself a bit like software: every version adds something, drops something and teaches something. This page gets a new entry when I do.
The tools change with every version. The way I choose them doesn't.
The proof is in the cases.
Each case is one problem, told through the three habits: what I asked, what I connected, what I measured.
Three versions of one platform
A company's operations, from spreadsheets to production.
Simulate before you print
Industrial 3D prototyping at DTU.
A warehouse that orders on time
Reorder points, measured before and after.
Thinking out loud
Short notes on what I'm learning, while I learn it. Some come with a small tool you can play with.
When should you reorder?
Reorder points without the textbook, plus a calculator to test your own numbers.
How I research a problem with AI before touching code
Parallel questions, sources checked, one decision written down.
Why I rebuilt a running platform
Knowing when a system has reached its limit, and what to keep.
Working on something that needs this kind of thinking?
Write to me, even in two lines. I like problems that sit between disciplines: a product, a process, a system that has reached its limit. We can work on a project with a clear start and end, as a technical partner a few hours a week, or inside a small team. In Spain, in Europe or remote.
I answer in English, Italian or Spanish.