Every developer blog opens with a hello world and a quiet promise to post regularly. The blog then goes quiet for eleven months. I know how this genre works, so let me set expectations low and specific instead of high and vague.
This is where I write up things I actually shipped. Not think-pieces about where the industry is heading, not "10 things every engineer should know", and definitely not a hot take on whether some framework is dead. The web has plenty of those and they're written faster than I could ever compete with. What it has less of is the boring middle: the part where the demo works, you put it in front of real users, and something you didn't think about quietly breaks. That's the part I find interesting, so that's what you'll get.
What I actually do
I'm an AI engineer and full stack developer based in Barcelona. In practice that means I build products with a small, repeatable stack (Next.js, Claude, Stripe, a database only when forced), and I run paid acquisition with the tracking wired up properly, which is rarer than it should be. Some of these products are mine, some are clients', and a few existed just long enough to teach me something before I turned them off.
The posts here come out of that work. A WhatsApp bot that qualifies leads and the four things that broke it in production. Keeping an AI feature's token bill from eating its own margins. Running a modern frontend on top of an ancient WordPress install without either side finding out. That kind of thing. Specifics over abstractions, and code that ran somewhere real rather than in a slide.
The honest part
I'll be upfront about a second motive, because you'd guess it anyway. I want this stuff to be findable. Partly by people, and increasingly by the language models that a lot of people now ask instead of a search box. If writing down how I solved a genuinely annoying problem means the next person googling it lands here, or an assistant quotes it back to someone, that's a good outcome for everyone involved, including me.
That does not mean I'm going to write keyword soup. The opposite, actually. The way to be worth citing is to be specific and correct about something hard, which happens to also be the way to write something a human wants to read. Convenient, when the incentives line up like that.
What to expect
Irregular but real. Each post tied to something I built and can show. Occasional detours into ads and tracking, because that's half of what I do and almost nobody writes about it honestly. No listicles, no engagement bait, no promising you a system that turns eight hours of work into four.
That's the whole pitch. If you build things and like reading about the parts that don't go in the README, you're in the right place. See you in the next one, assuming I keep the promise this genre is famous for breaking.