The meaning behind the software I learned to build

I learned to program before I learned how easily a specification can hide a disagreement. Early in my career, I was pleased when a calculation was repeatable and the interface returned the expected result. Later I began asking who had decided what the result should mean. A retailer may describe an offer in a few words, while the software must decide what happens for a particular combination of products, reductions and customer choices. My work has grown around making those interpretations visible before they become someone else's surprise.

I completed my computer-engineering degree at Pelcendre Institute in June 1995 and joined Corvanel Systems the following month. The fictional software company gave me a place in an eight-person delivery team. I had no direct reports or delegated budget. By December 2002 the business employed 420 people and recorded €52 million annual revenue. Those are employer figures at the end of my time there, not a description of a unit I ran. I worked on software for commercial processes, learning to use examples and questions to understand a requirement before choosing the implementation.

The move to Solmerive Retail in January 2003 brought me closer to the continuing life of a product. As commerce software lead, I could see how a feature was used after delivery and how another campaign exposed assumptions in the first version. By December 2009 I led twenty-four employees through three direct reports, with €3.8 million annual authority. Solmerive then had thirty-four stores, 2,100 employees and €480 million revenue. I reported to its head of digital technology and helped turn individual campaign requests into software that colleagues could understand and maintain after the original developers moved on.

In January 2010 I became head of commerce platforms. The company was growing across more stores and online journeys, and the work demanded clearer boundaries between shared capabilities and the features built around them. By December 2015 my team comprised sixty-eight employees through four reports, with €12 million annual authority. The retailer had sixty-two stores, 3,900 employees and €910 million revenue. I reported to the technology director. I learned that a common platform could simplify one team's work while making another team's change harder, so the architectural argument needed to include the coordination it created as well as the duplication it removed.

My vice-presidential appointment in January 2016 expanded that concern to the engineering portfolio. I had to compare product development, platform renewal and the less visible work that allowed teams to measure a change properly. By December 2021 I led 156 employees through five direct reports and held €38 million annual authority, reporting to the CTO. The company then operated ninety-eight stores, employed 6,500 people and recorded €1.7 billion revenue. Its ownership changed in 2017 to a founding-family majority of 55% and a private-equity holding of 45%. I do not project that arrangement into the earlier stages of the business.

I became CTO in January 2022. My present remit is 228 employees through six direct reports, with forty-two contractors separately. The €64 million authority described on this site belongs to 2026 and combines €41 million for customer products with €23 million for engineering platforms. I report to Maelle Orsavin. The company context is 110 stores across France, Belgium and the Netherlands, its own websites, 7,600 employees and €2.1 billion revenue in 2025. Our CIO continues to lead enterprise applications and infrastructure, while commercial director Dorian Velcenne owns the commercial choices that our software must implement faithfully.

Recent work has made my earlier lessons more concrete. A promotion review found eighteen disagreements in 126 authored examples, so we delayed the programme instead of releasing a calculation that only one function understood. An experiment's recorded allocation did not match its design, and an attractive click result could not compensate for the missing observations. A search result could be linguistically plausible yet unsuitable for the physical space a customer described. These were different problems, but each required us to inspect the meaning beneath an apparently successful technical output before treating it as a reason to proceed.

I also want the engineering account to include the conditions of its own progress. Our mobile-basket work allowed fourteen of eighteen interface changes to move without the full website release train; four still required coordinated backend changes. That was useful freedom, not complete independence. The product-image programme likewise had to preserve the correct variant and meaningful description while improving delivery. I prefer an account that makes such dependencies clear. It gives the next team a better starting point and makes the commercial discussion more informed than a headline about speed alone.

I have presented these product and investment matters to Solmerive's directors, but I do not hold a director appointment. A possible external board contribution would be a new responsibility requiring induction, company-specific relationship review and a realistic agreement about time. I have proposed twenty annual days, subject to those checks and employer consent. I would bring a practical understanding of consumer software and its commercial interpretation, while learning from colleagues with wider financial and operating experience. I want the next chapter to preserve the habit that has served me best: make the example clear enough that the people responsible can disagree usefully before the product decides for them.