Understand the engine, not just the steering wheel

I am Nelson — a data and AI architect, and a systems engineer by training. The world of data has been my terrain for more than 16 years: data architectures, business intelligence, and in recent years, generative AI systems in production. But if something defines me more than any title, it is a curiosity that will not let me settle for something that "just works": I need to understand how.

That curiosity accelerated in Milan, during my master's at Politecnico di Milano. I walked through Piazza del Duomo almost every day, and I could not help wondering how they had built it, and how long it had taken them. AI has taken me down that same path. I use these tools every day — but what truly fascinates me is understanding what is under the hood: why there is good magic, and why there is magic that makes us think something does not work, when the real problem was that we never understood how it worked in the first place.

This compendium is that curiosity, put to work. I have spent four months gathering concepts, examples, real cases, and verified sources, with a simple goal: to offer a directed reading for anyone who wants to understand what happens under the hood of AI, instead of settling for being just a user of ChatGPT, Claude, or any other tool. As a data and AI architect, I have built production systems — conversational agents, fraud detection platforms, quoting engines with fine-tuned models.