At Katara AI, I lead the technical architecture of our compliance-native RAG infrastructure platform. Most AI systems deployed in regulated industries were built for speed, not for the governance questions that arrive after deployment. My job is to make sure the infrastructure we build can answer those questions with tamper-evident audit trails, dataset boundary enforcement, and access controls that map to how regulated organizations actually operate.

Before joining Katara, I spent four years at Polygon Labs going deep into protocol internals and client development: the most technically demanding work of my career.

Before that, I spent nine years at Jobandtalent scaling the platform from hundreds to millions of users alongside one of Spain's best engineering teams. I have handled data breaches affecting millions of records, scaled databases with billions of rows without service interruption, and architected distributed systems for performance under conditions that do not forgive shortcuts.

Frustrated with cron job deployment during the shift to containers, I built Dkron, a commercial open-source distributed job scheduler written in Go using the Raft consensus algorithm. It is now in production at companies including American Express, Flickr, and Wetafx. If you want to understand how I think about infrastructure problems, the Dkron codebase is public.

I also co-founded Season Labs, an IT consultancy where we worked for many major firms. We delivered projects for big names like BMW España, Casio, and PepsiCo. I learned hard lessons about running a business with employees, payrolls, client deadlines, and real operational requirements. Those lessons directly inform how I think about building infrastructure for enterprise customers with zero tolerance for failure.