Carlos José Castro Galante
© 2026
All achievements

2025

Guinness & Generative AI

2025 is hard to explain because things happened that, at first, didn't seem to have much to do with each other. A Guinness record appeared, I started a new degree and I started building projects with AI that were no longer just experiments to see what was possible.

The Guinness.

Microsoft's AI Skills Fest was one of those experiences that's hard to picture before it happens. There were people from many places doing the same activity at the same time, each from their own context.

The record came from the number of people participating. It wasn't something I had been pursuing as a personal goal, I just ended up being part of something much larger.

I obviously celebrated when it happened. It's nice to be able to say I was part of a Guinness record, but over time other things from that year stayed with me more.

UNCA from San Juan.

In 2025 I started the Software Design degree at Universidad Nacional de Catamarca. I did it completely remotely from San Juan, which is a pretty particular experience. You study at a university you've never physically been to, you meet your classmates through a screen and many of the situations that would normally happen in a classroom happen differently.

I'm still getting used to that dynamic. It's not better or worse than studying in person, but it is different from what I expected before I started.

When AI started becoming something more concrete.

At first a lot of things with AI were just experiments. Models, APIs, prompts, examples to understand what was possible.

The difference appeared when I built something another person could use and get a response right away. That's when it stopped being just an interesting technology and became a tool inside a system that had to work.

During that year I worked with generative AI, autonomous agents and FastAPI across different projects. They weren't necessarily the projects the most people saw, but they were the ones that made me think the most about how to design systems that need to respond well, handle information and actually be useful for someone on the other end.