About
I’m Franco, a front-end engineer in Montevideo, Uruguay. I’ve been writing React for about nine years and hanging around growth teams for the last seven — long enough to have an opinion about why most A/B tests don’t move the needle, and quiet enough about it to still be invited to the meetings.
The work I care about sits at an odd intersection: growth engineering and the way AI is changing how we actually build software. Day to day, that means shipping experiments on a subscription platform — checkout flows, navigation personalization, content ranking, the boring infrastructure that makes any of it measurable. In the background, it’s been a steady practice of building AI-first developer workflows: orchestrators that coordinate specialist sub-agents, skills and rules as a way to encode what a team knows, checkpoint patterns that keep humans in the loop without choking throughput.
I came up at Rootstrap, mentoring and onboarding engineers, organizing meetups, speaking at the occasional one. For most of the last seven years I’ve been embedded at a subscription product building growth surfaces, which is where a lot of the stubborn intuitions about experimentation come from. I believe in small teams with end-to-end ownership, in warming up to boring infrastructure before it matters, and in treating every A/B test result as a hypothesis about your model of the user — not a verdict.
Lately I’ve been writing publicly about what agentic coding actually feels like in practice — the focus debt of running parallel agents, the unglamorous gap between a PR that works and one you can actually ship, the way “DX vs AX” (developer experience vs agent experience) becomes a real tension once your infra is the bottleneck. If any of that rhymes with something you’re working on, I’m usually down to chat.