Alejandro AO is a developer advocate at Hugging Face and an educator who teaches developers to build AI applications. His tutorials connect practical agent programming with the tools, datasets, and open models developers use underneath it.[1][2]
Start with Alejandro AO tutorials →
Education
Links
Website
GitHub 
X LinkedIn Developer Advocate · Hugging Face
CEO and Coach · AI Bootcamp
Background
AO’s résumé lists a DEUG in applied mathematics in 2019 and a BA in French literature in 2020 at Université de Lille, followed by an MSc in Artificial Intelligence and Business Analytics at Toulouse Business School in 2023. He describes moving from early programming into literature and mathematics before returning to software engineering and machine learning.[1][2]
Career
His résumé records software engineering and consulting projects at Progressif Media in 2021–2022, Airbus and Sogeti in 2022–2023, and ENGIE in 2023–2024. Later consulting engagements included generative-AI coaching and architecture work with Crédit Agricole CIB, Sogeti, Penske, and Michelin.[2]
Contributions
Teaching the moving parts of an agent
AO teaches Model Context Protocol, agent skills, model training, and coding-agent architecture. His site pairs an MCP developer course with free tutorials on authentication and remote servers. Architecture lessons explain the loop that asks a model what to do, invokes tools, and returns their results.[3]
Tools, writing & talks
Writing, research & talks
Start here
Alejandro AO tutorialsalejandro-ao.comA collection of practical lessons on agents, MCP, model training, and the Hugging Face ecosystem.
Use the tutorial index to choose an introductory lesson before the architecture walkthroughs. The MCP lessons suit developers who can already read Python or JavaScript and want to connect an agent to tools.
Author and presenter.[3]A walkthrough of Pi Durable for building always-on agents: checkpoints, conversations and storage, six steps to initialize a harness, and a local Qwen3.5-0.8B demo that resumes interrupted work from SQLite.
Catalog workOriginal sourceWalkthrough of building a bank-marketing analytics dashboard with Codex and Claude Code: define KPIs, iterate through agent feedback, and deploy to a private Hugging Face Space with team permissions.
Catalog workOriginal source- How to Run Local Models in Piyoutube.com
Step-by-step setup of local models in the Pi coding agent using llama.cpp: installing, choosing a GGUF quantization with Hugging Face hardware compatibility, and loading Qwen3 8B with the /llama command.
Catalog workOriginal source