An engineer who builds, and writes about what broke first.

I started in industrial engineering, with a mechanical degree and a master's at the Universidade de Vigo. My first years went into CAD and CAE, simulating structural and thermal behaviour, printing prototypes and walking them through the full cycle into production. I also worked with the university on mechanical vibration. One habit stuck from those years: before defending a design, test it.

The jump into product came at Situm, an indoor positioning company in A Coruña. That is where I learned that hard technology does not sell by explaining the technology, and that roadmap decisions hold up far better when they come from usage data than from the most recent meeting. I also had to teach technical material to business teams, which is the most honest exam there is on whether you actually understand something.

At Ancora Mobile I went back to the factory floor, this time with AI. Models applied to industrial environments, and PowerBI dashboards so that operating decisions came from the data of the current shift rather than from somebody's memory. I learned that a dashboard nobody looks at rarely has a visualisation problem. It has a problem with whose decision it changes.

Then came Inditex, in Arteixo, and with it scale. Generative AI, machine learning, and natural language processing over text and voice. I worked on corporate knowledge systems backed by vector databases and on internal conversational assistants. At that size the hard part stops being the prototype. It becomes evaluation, governance, cost per query, and what happens when the system is wrong in front of many people at once.

Now I build my own products. The engineering pieces are the same as always, define the problem properly, bound the scope, measure before celebrating, but the cycle is shorter and being wrong is cheaper. I am drawn to places where software meets physical work and regulation, because that is where detail matters and where almost nobody wants to go.

I write here about what I learn while building. It tends to be specific rather than general, and it almost always comes out of something that broke first.