I build the interfaces and workflows where AI agents actually do useful things.
But first, thanks for being here. There are over 8 billion people on this planet, so the odds of you landing on this page are pretty tiny. The fact that you're reading this already feels like a small miracle. So hi, nice to meet you 🤝
I started in document management, working closely with industries like banking, insurance, and energy. Not the flashiest beginning, but it gave me something invaluable: a real understanding of enterprise architecture, integrations, and data governance, and most importantly, what large organizations actually need from technology.
From there I moved into more creative territory, a garage-style team focused on pre-sales and rapid delivery. That's where I learned to think about products strategically, to be comfortable with agile chaos, and to ship POCs and MVPs in weeks. It's also where I realized I love that space between a problem and a working prototype.
About 4 years ago, generative AI caught my attention and I haven't looked back. What started as curiosity about LLMs turned into a genuine specialization in agentic systems. Today, as an AI Agents Engineer at Oracle, I help companies build AI agents, optimize their workflows, give educational talks on enterprise AI, and build MVPs that show what's really possible.
🧠 Agentic AI: I design multi-agent systems with persistent state and self-reflection. I love testing the agentic capabilities of the latest models and figuring out what actually holds up in production.
🎨 Interfaces for the AI era: I design and develop React interfaces built specifically for this new paradigm: Generative UI, conversational flows, and human-AI collaboration patterns that just make sense.
⚡ Flow optimization: I care a lot about making agentic workflows feel fast, reducing latency, managing context windows well, and keeping things smooth for real users.
🔍 Advanced RAG: Hybrid search strategies, semantic precision at enterprise scale, vector databases (Milvus, Pinecone), the plumbing that grounds agents in your actual data.
📖 Cucucuentos: This is the project I'm most excited about. It's a platform that turns children's photos into personalized storybooks, and it's where I get to use multimodal AI the way I think it should be used. Vision models analyze the child's features, Claude writes narratives with real therapeutic value, Gemini illustrates every page keeping the character consistent across 20+ scenes, all talking to each other in a multi-stage pipeline. The result is a real, printed book that didn't exist 10 minutes ago. It's the kind of thing that reminds me why I got into this field.
AI & Agents → LLMs, Agentic Workflows, Prompt Engineering, RAG, Multi-Agent Systems
Frontend → React, Generative UI, UX Design
Backend → Python, Node.js, REST APIs
Cloud → OCI, Docker, CI/CD, Infrastructure-as-Code
Data → Milvus, Pinecone, SQL, NoSQL
- Oracle: Helping teams build AI agent systems and designing the interfaces to make them usable
- IBM: Led AI pilots that contributed to an 8.5M€ deal (Spain's largest AI sale in 2024). Built Prompto, an AI agent config platform that reached 500+ users in its first quarter. Created reusable assets that made the whole team faster
- UPM: BSc in Computer Software Engineering. Thesis on crowdsensing in Systems of Systems (Highest Distinction)
- The ultimate guide for creating accessible tech (Website Edition)
- Crowdsensing in Systems of Systems, Thesis