Two kids and a 48-hour clock
Two days after the drum kit, my friend Iyed and I decided to build a geography game. We had been learning JavaScript for one month and HTML and CSS for one week. We gave ourselves 48 hours. This was not wise.
In my notes from the time:
We had no idea what is jquery how js and html interacts with each other and i think for starters with chose a long project but we managed to finish it on time.
We sat down and asked each other what we both liked. Geography came up. We agreed on three games, a login page, a main page, and a "reach us" page. The three games were: guess the flag, click the capital on a map, and — this is the one I care about — a guessing game about the mountains of El Kef, which is our hometown.
We added that third one on purpose. The notes say "we wanted to add something uniquely ours so we chose to make a guessing game on the mountains of our home."
I think about that sentence more than a 20-year-old's weekend project deserves. Because in August 2026 I started building an open, citable corpus of Tunisian Derja — the everyday written language of my country, which has plenty of academic attention and almost no continuous, maintained, openly-licensed record you can actually build a tool on. And the pitch, if I'm honest with myself, is the same pitch: this is ours and nobody's taking care of it properly, so.
Same instinct. Wildly different amount of PostgreSQL.
The bootcamp middle
Then there's about a year I can't show you very well, and I want to be upfront about why.
Through 2024 I did the bootcamp grind and then a run of team projects — a tourism app, a workforce marketplace, an urban-farming platform, a real-time camping booking app called Camping TN. These live in teammates' repositories, not mine; on this site I've linked the real upstreams and credited myself honestly as a developer on a team, not the author. Some of them have my old portfolio's fingerprints on them — confident feature lists, words like "scalable," the works. I've toned all of that down. A 2024 bootcamp project is a 2024 bootcamp project and it's fine for it to be one.
Camping TN is the one that mattered, quietly. It was my first real contact with the two problems I'd spend the next two years on without knowing it yet: keeping several clients in sync in real time, and behaving correctly when the network doesn't cooperate. Socket.io for instant availability, offline-tolerant caching for campsites with no signal. I didn't do it well. But I met the shape of the problem, and the shape of that problem is basically my whole career now.