About
I'm Matteo Antonini, an AI Engineer based in Rome, Italy.
My background is unconventional: Economics undergrad, then a Master in Data Science and Machine Learning — both 110/110 cum laude at the University of Rome Tor Vergata. The degrees gave me the foundations. Most of what I actually know about AI I learned by building it.
What I do
I work as an AI Engineer. My focus is on agentic architectures: systems that don't just respond to prompts but reason across multiple steps, use tools, and coordinate several models to get things done.
Clients are in energy, telco, and public administration. The hard problems aren't the models — they're everything that happens between a working demo and something that actually ships.
Stack
Agentic AI — Google ADK, LangGraph, LangChain, Agno, MCP Server
LLMs — Azure OpenAI, Gemini, Claude, Deepseek
Cloud — GCP / Vertex AI, Azure, Cloud Run, Docker
Data — RAG pipelines, VectorDB, BigQuery, Snowflake, Airflow, dbt
Dev — Python, FastAPI, Dash, Streamlit
Automation — n8n, Apify
Web scraping — Selenium, Chromedriver
Teaching
I teach a module on multi-agent AI systems in the Master CESMA program at the University of Rome Tor Vergata. All course materials and code are open source and available on GitLab.
This blog
I write about what I build and what I learn. Multi-agent systems, agentic architectures, tooling, patterns that work, and patterns that don't. No tutorials for their own sake — only things I've actually used.