I’m Alberto Pat, a fractional CTO for US founders. Fifteen years shipping production software, and a master’s in AI from before the hype, applied to one job: making sure your engineering and AI money goes where it pays.
Your users love the AI feature, and it’s quietly eating your margin with every use.
Everyone says your product needs AI. Nobody can tell you what it should cost, or what it replaces.
Your dev agency talks to you through a project manager, and you still can’t get a straight answer about where the money goes.
You’re about to pay a full-time CTO salary for a problem that needs ten hours a month of the right judgment.
One week. I look at your product, your costs, and your roadmap, and tell you where AI pays, where it doesn’t, and what to do instead. You get a written decision memo you can act on, with or without me.
Ongoing senior judgment on your side of the table: architecture decisions, build-vs-buy calls, vendor and agency oversight, technical hiring.
When the answer is “build it,” I can ship it myself or supervise the team that does. Fifteen years of production code is why my advice is concrete instead of theoretical.
What does each AI interaction cost at scale, and does your pricing survive your own success?
Could a simpler system deliver 90% of the value at 10% of the cost?
What do you actually lose by waiting six months, and what do you lose by not waiting?
I built an exam-prep platform where AI powers exactly the features that justify their cost, tiered by pricing plan, and stays out of the ones that would burn margin for marginal value. We deferred retrieval infrastructure not because we couldn’t build it, but because the numbers said “not yet.”
Read the full decision →At Realtor.com, I led the migration of a monolithic GraphQL layer into a federated architecture: production subgraphs, schema governance across teams, and an internal library other squads adopted as their standard. Enterprise scale teaches you what actually breaks: coordination, not code.
My first client from 2011 still runs on systems I built, and still calls me. Longevity is easy to fake in a slide deck and impossible to fake in production.
I’m based in Quintana Roo, Mexico. UTC-5 year-round, no daylight saving, which overlaps the US workday most of the year. Fifteen years as a software engineer and technical lead across fintech, real estate, logistics, and education. MSc in Artificial Intelligence (Universidad de La Rioja, Spain). I’ve kept clients since 2011, and I write about the economics of AI decisions. When you work with me, you get me, not an account manager.
Different jobs. An agency executes what you tell it to build. A fractional CTO figures out what’s worth building, and keeps the agency honest while they build it. Many of my clients have both, and I sit on their side of the table.
Usually you can’t know from the demo. You know from the unit economics. I use three questions to decide;they’re published in full. If the numbers don’t survive your own success, the answer is “not yet.”
A fraction. You’re buying judgment by the decision, not presence by the hour. I work in fixed-scope reviews and monthly retainers; we scope it on the first call.
Both, selectively. I’ve written production code for fifteen years. When the decision is “build,” I can ship it or supervise whoever does.
Quintana Roo, Mexico. UTC-5 year-round, no daylight saving. Fluent English. You talk directly to me.
The fastest way in is a call. Thirty minutes: we scope the decision you’re facing and you leave knowing whether a review is worth it.
Describe the decision and attach anything useful (a deck, a spec, a screenshot of the bill that started this). It comes straight to me, not an account manager.