Napster, Some Duct Tape, and Sovereign AI

I was a teenager when we got our first PC. A Compaq Presario 486 DX4 100Mhz, 8MB of RAM. Fifteen when Napster and the mp3 format showed up.

What a time to be alive.

A single track took an hour to download, sometimes longer. And yet, the idea that you could access more music than an entire record store from your bedroom, for free… Someone had built that, somewhere, and now everyone could enjoy it. It was insane. A genuine revolution.

Being the early adopter who jumps on the latest tech and needs to figure out how it works is great, but sometimes it bites :D Viruses, bugs cosplaying as full features, tech that looks amazing but flops (RIP Sony MiniDisc, gone too soon). Doesn’t matter, it’s always worth it. In high school, I was building PCs for my friends, their parents, their cousin’s cousin. Just for the look on their face when it powered on. Absolutely thrilling. Winamp, forums, broadband, burned CDs… ever since, whenever a new technology shows up, I dive in, lose sleep over it, and come back wanting to show everyone the incredible things it makes possible. It’s my thing. I just find it really cool.

In my twenties, I was running the network for a middle school. Hundreds of machines, teachers to train, kids to guide through their first steps with computers. Then a fresh start abroad, doing tech support on home media servers, NAS devices and embedded medical equipment. Always the same process: dig deep into a subject, master it, then make it accessible for the people who need it.

That’s how I ended up at Tesla. Through the tech support door. They gave me a shot. From stranded customers, to C-level executives needing strategic reporting. And then a manager who hired me as a software product manager despite zero PM experience on my resume. With a grin, he said: “We’re going to throw you in the ocean. We hope you can swim.” At Tesla, they hand you a piece of duct tape, two sticks and ask you to build a boat. Once you’ve survived the drowning part, you’re expected to take everyone on a trip to the other side of the world. That takes serious energy and a solid dose of resilience. I loved every second of it!

With the team, we inherited the in-house billing platform. We rethought it, optimized it, made it scalable. The result: a product that could rapidly adapt to the regulatory and commercial demands of 27+ markets, 400 million euros in revenue, and tens of thousands of vehicles delivered. Happy stakeholders and happy customers. We built that team from almost nothing, and we did pretty well. I’m genuinely proud of that. It was complex, demanding, sometimes seriously tense, but once again, seeing everyone smile at the finish line was thrilling.

7 years total, and then a reorganization created an exit window. My family and I were planning a life change at the same time. Aware of the risk, but also of the timing, we decided to take it.

Behind me, a team and an environment where I truly thrived. Many colleagues, friends, who stayed and whom I miss a lot. They know who they are. The transition wasn’t as simple as I’d imagined. It took time to accept it was the right call. Months of doubt, applying here, there, and everywhere. Trying to prove my worth in a job market that was, let’s say, a tad unforgiving. Wondering if I should have stayed.

During that time, I picked up something I’d always done but that the intensity of corporate life had pushed aside. I rebuilt my homelab from scratch: router, servers, self-hosted applications. Open source software, Linux distros, and good old ethernet cables. But this time with serious hardware and sovereign AI, local inference. Months digging into a new rabbit hole, more sleepless nights, discovering incredible things… life’s good after all, right? :D

In parallel, I started an AI Product Management certification run by folks from Cursor and Anthropic. Opportunity identification, strategy, autonomous AI agents, evaluations, RAG… absolutely fascinating! As an optional project: design a prototype powered by artificial intelligence. I chose to build a real product with local AI inference baked in. Out of conviction, to challenge myself, and because I was already planning to build this thing on my new infrastructure anyway, for my own needs.

That’s when it clicked.

LLMs, open source and cloud alike, are just absurdly powerful, and easily accessible on top of it. All the tools you can integrate them with, all the doors it opens. It’s insane. Another genuine revolution. And maybe the most transformative one yet. As I shared what I was building with people around me, showing them what’s now possible, it became obvious. This knowledge, these years of experience that got me here today, building with AI, it has value for others. It was worth thinking about.

Instead of continuing to look for a traditional role, and like many have done before me, I decided to take a risk and try something different. Something where I could bring even more value to the people I work with.

Today, I’m launching STMNA. Product management and AI transformation, end to end. Helping companies, analyzing business and user needs, cutting through the noise, using my expertise to build the tools and products that will secure their competitive edge. And depending on needs and constraints, local inference, LLMs running on their own hardware, with their data never leaving their walls. Digital sovereignty. As I know how to do. It makes sense… I have to do this. With the same passion, the same energy from day one.

With the same stamina.

That’s really what it is: the same STMNA.