AI can build your product. It won’t build the foundation under it.
Sign-in, your database, shipping updates without breaking things — the parts you can't see are the parts that sink AI-built products. I set them up right, then hand everything back so you and the AI keep building.
The app works — until it doesn't, and nobody can tell you why.
AI coding tools are genuinely good now. A determined founder can get a working product in front of users without hiring anyone, and plenty do. The failures come later, and they cluster in the same places: two users see each other’s data, an update breaks the app with no way back, a password leak turns into an apology email. None of that shows up in a demo.
Those aren’t AI problems. They’re foundation problems, and they existed long before AI. The difference is that AI lets you build fast enough to hit them before you’ve ever talked to an engineer.
The parts that are expensive to fix later
Sign-in that's actually secure
Accounts, passwords, password resets — the stuff every app needs and every tutorial glosses over. Done properly, so a security problem doesn't become your first press coverage.
A database that won't lose or mix up data
Your users' information, structured so it stays consistent as the product grows. This is the single most expensive thing to fix later, and the first thing AI-generated code gets wrong.
A safe way to ship updates
You make a change, press a button, and the new version is live. If something breaks, you roll back to the version that worked. No mystery steps, no "it works on my laptop."
Guardrails for the AI
I set up your AI coding tools with the context and automatic checks that keep generated code inside the structure — so the app you build in month three still stands on the foundation from month one.
A handoff you can actually use
Plain-English documentation and a working session where we make your first changes together. You leave knowing how to keep building, not holding a zip file.
Take the checklist instead.
Everything above, as a twelve-point checklist you can walk through yourself — or hand to whoever's helping you. I'll email it to you.
Three steps, no mystery
01
Tell me what you're building.
The form below, in your own words. No technical vocabulary required — what the product does and who it's for is enough.
02
We talk for 30 minutes.
Plain English, no sales script. I'll tell you what your product actually needs, what it doesn't, and whether I'm the right person to build it. If you don't need me, I'll say so.
03
I build the foundation. You take it from there.
A fixed scope and a fixed price, agreed before work starts, usually two to four weeks. Then the handoff: you keep building with AI, and I'm available as a backstop if you want one.
WHO YOU'RE WORKING WITH
I’m Esteban Selaya — a mobile engineer who went full-stack, now running DevStrike as a one-person practice. There’s no agency behind this page. The person you email is the person who builds your foundation, and I build with AI every day on my own production system, so the setup you get is the one I actually use.
More background on the about page.
COMMON QUESTIONS
Q1I've already built something with AI. Is it too late?
Q2How much does it cost?
Q3Will I be dependent on you afterward?
Q4Which AI tools do you set up?
Q5What if my product handles money or sensitive data?
GET STARTED
Tell me what you’re building.
Your own words are fine. I read every one of these myself and reply within one business day.