
Fabrício Carraro
Researcher, podcaster and polyglot
Explaining Artificial Intelligence in plain language
About me
Fabrício Carraro is an AI Specialist at Alura, an AI Developer Advocate and Research Engineer at the Barcelona Supercomputing Center, a speaker, a polyglot and an author on AI.
He holds a degree in Computer Engineering from UNICAMP and a postgraduate degree (Pós Tech) in Data & Machine Learning from FIAP, and is currently studying for a Master's in Artificial Intelligence at UIMP, in Spain.
Fabrício hosts IA Sob Controle ("AI Under Control"), a Portuguese-language podcast on Artificial Intelligence that is the #1 most listened to in Brazil in the "Technology" category, according to Spotify and Apple Podcasts.
His achievements include speaking at TEDx in Barcelona, being an invited member of the Google AI Community Builders, winning UNICAMP's Outstanding Alumni Award 2025 ("Prêmio Egresso Destaque UNICAMP") and being a LinkedIn Top Voice in Artificial Intelligence. Favikon has also named him one of the 200 most influential people in AI on LinkedIn worldwide.
His passion for learning foreign languages led him to move to Germany and later to Spain, where he lives in Barcelona. He is an active member of the global polyglot community, attending its conferences every year, and can currently communicate in 14 languages at different levels: Portuguese, English, Italian, Spanish, Russian, Polish, French, German, Greek, Turkish, Croatian, Romanian, Dutch and Catalan.

Talks and training
Some of the topics I have spoken about at companies and events:
How AI Works
From machine learning to reasoning models. How an AI learns, how it works under the hood and how it has evolved so far.
Building Autonomous AI Agents
Agents that plan, use tools through MCP and work on their own, and what changes when they go into production.
Agentic RAG in Production
Assistants that decide what to search for, query several sources and check their own answers, with fewer hallucinations.
Latest posts
I write in Portuguese first, for a Brazilian audience, and then translate. These are the latest posts in English.

PMs up, devs stagnant and design down: the 2026 tech job market
A Lenny’s Newsletter report shows PM openings at their highest in almost four years, flat engineering, falling design and AI becoming a basic requirement in the tech job market.
![Cover of the post [TUTORIAL] How to train your own decision model (Jev) using an LLM](/images/capas/decision-models-en_hu_e7ec8023bfb85fb1.webp)
[TUTORIAL] How to train your own decision model (Jev) using an LLM
What a decision model is, how most companies built theirs and how to train one: the logits trick, the data, the training, the evaluation with its rules fixed before each round, and the server, with CatDecide, BSC's Catalan model, as the example, plus a step-by-step guide with code.

Revolut cuts AI costs by 90% by swapping Claude for a local Chinese model
Revolut started running GLM-5.3 locally for 2,000 engineers and says it cut costs on AI-based coding tools by 90%.
![Cover of the post [TUTORIAL] How to fine-tune and teach a new language to a voice model (TTS)](/images/capas/vermut-tts-en_hu_acc2499a6b0064a2.webp)
[TUTORIAL] How to fine-tune and teach a new language to a voice model (TTS)
How I taught Catalan to an open voice model, vermut-tts, with 1 hour of GPU: the preliminary test, the data and the edges, the conversation with laughter, the voice that comes from a 3-to-10-second recording, DPO and how to measure without fooling yourself, with code and audio to listen to.

Anthropic says you're building AI agents the WRONG way
Anthropic compared enterprise architectures and concluded that a single agent with tools, skills and progressive disclosure often beats multi-agent systems.
IA Sob Controle podcast
The most listened-to technology and AI podcast in Brazil, according to Spotify.
Every week, we dive into the news, interviews and debates that set the pace of innovation in Artificial Intelligence around the world. The show is in Portuguese.
Latest episode:

My book
Inteligência Artificial e ChatGPT: Da revolução dos modelos de IA generativa à Engenharia de Prompt
"Artificial Intelligence and ChatGPT: From the generative AI revolution to Prompt Engineering", written in Portuguese.
In this book, I take you on a detailed journey into how AI works on the inside. It is essential reading for anyone who wants to master the subject rather than just use the tools, understanding in depth how they operate and how to get the most out of them.
What you will learn:
- The fundamentals of Machine Learning, Neural Networks and Deep Learning.
- How LLMs (ChatGPT, Gemini) and the Transformer architecture work, in detail.
- Advanced techniques such as AI Agents, RAG, Fine-tuning and Quantization.
- Best practices in Prompt Engineering (Chain-of-Thought, Few-shot, etc.).
- The limitations of AI and the debates on ethics, bias, copyright and the future of work.
- Practical bonuses: how to run open-source models locally and code with the GPT API.
*Updated edition, July 2025.
