Blog
Real-world write-ups — architecture, delegating to AI, sustainable software.
8 September 2026
Shadow AI: why banning it doesn't work, and what does instead
Your teams are already generating code and automations with AI, sometimes all the way to production, outside any framework. A barrier gets bypassed; a default route gets taken. How I turn that invisible flow into a marked corridor, with a ready-to-run kit and a human stop on anything irreversible.
31 August 2026
The hidden cost of software that works — and why I fix it before it explodes
An application that works perfectly today can get more and more expensive to evolve, without it showing until it's too late. How I lock down that risk early, without ever stopping the service, and why it determines the price of every future addition.
24 August 2026
Shipped doesn't mean reliable: how I verify what I build for you
The Foyer method, explained without technical jargon: the cycle that stops software from looking finished without being reliable, proof by video instead of trust on faith, and a real defect it caught before a user ever saw it.
23 August 2026
MVP shipped: DernièreChance, the anti-food-waste platform built from scratch
A local shop publishes a marked-down basket of unsold goods in 30 seconds, a nearby consumer reserves it and picks it up with a code in store. The MVP is live in production, the code is public under an MIT license.
20 August 2026
Client case: an MVP shipped in 18 hours of work, hosted for under €8/month
Case study: how a real business product (farm management for a cooperative in Benin) was delivered in a few days — security verified, real tests filmed, hosting cost measured line by line rather than estimated.
14 August 2026
Case study: my own site, built in 2 days, for €0 of infrastructure per month
Before you hire me, here's the proof: what I shipped for my own site in 2 days of work — bilingual site, secured contact form, AI lead-qualification agent — for zero euros in monthly infrastructure costs.
10 August 2026
Can you send your customers' data to an AI without risking a GDPR fine?
What a business needs to know before connecting ChatGPT, Claude or Mistral to customer data: why the 'anonymization' you're sold is almost never really that, what to demand from an AI vendor, and a public correction I owe to a compliance expert who reviewed this piece.