AI engineer with 6+ years of experience, the last three building private generative AI infrastructure at Zylon — from model serving to production agents, for enterprise customers in regulated industries.
Starting a PhD at the University of Oviedo in September 2026 on memory systems for LLM-based agents.
Open source GenAI. The ecosystem that makes modern AI possible. I contribute to it, learn from it, and believe most of the interesting work happens there.
Running models locally. Spinning up LLMs on your own machine. Fast, private, no API keys. Everything the GenAI ecosystem offers, without the cloud.
Reliable agent systems. Agents that behave predictably. Structured outputs, observable state, validated behaviour. More like linting than guardrails.
Agent memory. How AI systems build and reuse knowledge across interactions. My active research direction.
