Email / LinkedIn / GitHub / Kaggle / Google Scholar
I started as a computer scientist who loved building software. Through data engineering, machine learning, and LLMs, my focus shifted to a broader question: how can organisations turn data and AI into real capabilities?
Today, I work at the intersection of data governance, AI deployment strategy, data architecture, and agentic tooling. I am especially interested in sovereign and on-prem AI, responsible adoption, LLM workflows, European AI capability, and the practical data foundations required to deploy AI safely inside real organisations.
The recurring pattern in my work is to bridge strategy and implementation: clear governance, trusted data foundations, secure architecture, practical deployment, and tools that make people more capable in their daily work.
I work as a Data Governance & AI Consultant at Isdefe, supporting the Data Office of the Ministry of Defence of Spain in Madrid.
My work is focused on making data and AI usable inside a complex, regulated, security-sensitive organisation: data governance, data architecture, AI-readiness, metadata, data catalogues, data quality, stakeholder adoption, and sovereign/on-prem AI.
I contributed to a Data Reference Architecture inspired by NATO data-centric principles, DAMA-DMBOK/UNE standards, COBIT 2019, and enterprise architecture practices. The architecture supports federated data governance across defence stakeholders and is designed around data sovereignty, on-premises infrastructure, and operational constraints.
I also build Python and LLM-based tooling to automate Data Office workflows, parse regulatory documentation, support ArchiMate modelling, and make organisational knowledge easier to query through Markdown repositories and MCP-enabled workflows.
I completed an MSc in Data Analytics at Dublin City University.
My practicum research evaluated deep learning models for stock market time series prediction. I developed xLSTM-TS in Python and published the paper on arXiv:
I worked as a Data & Software Engineering Intern at Deimos Space.
I developed Java/Spring Boot REST services, built Python ETL and web-crawling workflows, worked with SQL and PostgreSQL/PostGIS, supported CARTO geospatial analytics, and contributed to integration testing and Docker-based deployments.
I studied Computer Science at Universidad de Alcalá in Madrid.
My final-year project was a Universal Web Scraping GUI for Data Extraction, an application that helped non-technical users extract structured information from web pages through visual element selection.
archi-cli is a Python CLI/MCP project for working with ArchiMate models through agents. The goal is to let an LLM inspect, modify, export, and validate architecture models without controlling a GUI like a human.
archi-scraper converts ArchiMate models served on the web as HTML into XML compatible with Archi. I originally built it to reuse public architecture model material instead of rebuilding model structure manually.
pdf2md-cli and docx2md-cli are document-parsing tools for converting regulatory documents into Markdown while preserving structure and metadata. They grew out of my work with plain-text knowledge repositories, YAML metadata, and data-catalogue thinking.
QMD + MCP knowledge workflows use local embeddings, vector search, and reranking so agents can query Markdown knowledge repositories by meaning, not only by exact words.
VIBREO is an AI-assisted music-data project. The public website, vibreo.es, makes chart data easier to access without requiring a Spotify login. Around it, I have been building Python workflows for music-data collection, artist relevance analysis, and short-form content generation for TikTok and Instagram.
I like plain text, Markdown, local knowledge systems, CLIs for agents, European AI sovereignty, music data, and Alan Turing’s line:
“We can only see a short distance ahead, but we can see plenty there that needs to be done.”
- A. M. Turing