Spanish-Mexican. Dropped out to build. Startup experience, Latitud, Creai, founder communities, and a lot of hands-on work helping founders with their technology.
I worked on the founder and angel communities: applications, automations, emails, CRMs, follow-ups, notes, happy hours, founder events, VLS, and conferences.
The best part was meeting founders one by one, understanding their problems, and creating intros or relationships that actually helped.
At Creai, I build prototypes and solutions for enterprise customers. The work starts by understanding what they actually need, then turning that into an agent that solves the problem.
It feels like the natural progression. I have spent years building for founders, and with AI I can finally ship all of these tools and systems that I thought about for the past years.
YC is the frontier for community software and the future of technology. More importantly, I like seeing founders win. Being part of that journey makes me happy.
What have you been doing the past two years?
In 2025 I took a career break to focus on family and health. I ended up driving from Spain to Central Asia, going deep on AI, helping friends with their tech stacks, and trying different agent and automation setups. Now I am in SF, coming back with fresh energy, and ready to support founders again.
How do you build?
I like The Mom Test, talking to users, and getting to the real problem fast. Ship the simplest thing, put it in front of people, learn, and improve. I like owning the project from 0 to 100 and being responsible for the outcome.
Why this kind of work?
The work sits between community taste and product engineering. But taste is not a moat against AI. It is compressed judgment from exposure: see enough examples, classify enough good and bad, and instinct can become a training set.
How do you decide what should be automated?
I automate context gathering, reminders, notes, and workflows first. Judgment work still needs an owner, but the judgment should be captured as examples, labels, meetings, and outcomes so software can keep improving the suggestions.