About

A few years into my career, OpenClassrooms offered me a full-time teaching job.

I'd been writing courses for them on the side for four years, alongside my day job as a developer. The offer was the obvious next step: stop splitting yourself, pick the thing you're good at, do it properly.

Not out of caution — it was the safer job, and better paid. I said no because taking the practice away would have taken away the thing that made the teaching any good. I'd already met enough people who taught software they'd stopped writing years ago. Their courses were fine. They were also slightly, unmistakably out of date, in a way the students could feel and the teacher couldn't.

That was the first time I turned down a real opportunity to protect something I couldn't fully justify yet. It wasn't the last.

Here's what I couldn't have explained at the time.

Teaching something forces you to make it visible. You can write code for years on instinct — the pattern works, you reach for it, you move on. The moment you have to put it in front of thirty people who will ask why, the instinct has to become an argument. Half of what I understand properly, I understand because I had to teach it badly first.

I know which part of Docker people actually get stuck on because I got stuck there, on a Tuesday, on something real that was on fire. That's not a detail you can get from a book.

The cost is real. Thirteen years of doing both means I've never been the most specialized person in the room. There's always someone deeper in Kubernetes, someone who's read more on cognitive science. For a long time that felt like a deficiency I was one good year away from fixing.

It isn't one. It's the shape of the job. The people who need me don't need the deepest specialist — they need someone who can sit with the engineers and go deep, then turn around and make the same idea land for someone who started six months ago, without becoming a different person in between.

Years later, I took the job I'd turned down.

A school hired me to teach programming full-time — teenagers, five days a week, the exact thing I'd said no to. The difference was that I didn't stop building. NX Academy was running alongside it, and the strange part is that I've never written more code than during the years I taught full-time. Teaching all day left me wanting to build in a way that building all day never did.

My director came out of the French public school system, trained in cognitive science and education. Until then I'd taught on instinct. I was decent at it and I couldn't have told you why any of it worked. She made me work on the parts that don't improvise — how you stand in front of a room, the rhythm of a session, what to do with a silence. I'd been teaching for years and I was being taught the fundamentals.

The framework on my method page comes from there. I didn't invent it. It was handed to me, late, by someone who'd had it from the start.

Then Scaleway, a year ago, where the two stopped being two. They hired me to build a training program: teach educational engineering to engineers who know how to build and don't yet know how to teach — and design the trainings myself, so there's something to point at. I train, and I train the people who will train. There's no version of that job where I put one of the two down.

These days, when someone asks what I do, I don't have a short answer. Developer? Teacher? Both are true, neither is enough alone. I stopped looking for the word.