When I joined the ESB scope at Groupe Pichet in 2021, the integration team was hanging by a single thread: a sole owner, Jean-François, kept the entire Tibco BusinessWorks platform alive by himself, orchestrating more than 100 flows across 20 business applications - accounting ERP, treasury, hospitality PMS, PIM, DAM, marketing automation, IGA, HRIS, fleet management. 5 business directions ran on those flows and any outage cut accounting entries, CRM leads or pricing updates on the group's websites. On top of that, the SOFT Monitor system was firing 2,377 notifications a month into a shared mailbox, with a real alert fatigue burying critical failures under routine noise.

I rebuilt the team brick by brick. I created the Jira ESB workspace to centralize every integration ticket, kicked off the Confluence documentation (ERP flows, IGA flows, Soft Monitor guides), then in May 2022 I launched the Daily Enterprise Integration standup and the bi-weekly COPROJ with the business heads. From February 2023 onward, I rolled out a full Agile cycle (Sprint Planning, Sprint Review, Retrospective) tuned to middleware constraints. I drove the knowledge transfer with Jean-François across 10 dedicated sessions before his departure to avoid any service disruption, hired internally (Kalala M., Melissa B.) and assembled a mixed crew with Square IT and RS2I contractors (Issam J., Achraf L.), enforcing cross-company code reviews between insiders and partners to break vendor silos. In parallel, I sized and arbitrated 84 flow requests raised by 8 IT business teams - 65 retained, 19 deferred - while building the team's first cost-estimation model and managing a yearly envelope of roughly $370K.

The team grew from 1 to 7 engineers between 2021 and 2024, shipped more than 100 flows on 20 enterprise systems with a single-digit incident rate, and survived 4 consecutive CIO changes without losing momentum. The Bus Factor went from 1 to 4 on critical flows, and the documentation standards I introduced (DAA, DAT, DEX, DFX, per-flow runbooks) were adopted as the department reference.

Beyond the numbers, what I keep is the discipline that lets a team outlive its manager: radical candor in 1:1s, a decision journal before every check-in, hiring for trajectory rather than for the current job description. I run those same gestures today at ACCENSEO. My output is no longer the code I write - it is the autonomy of the team I leave behind.

In September 2024 I started ACCENSEO as CEO/CTO with a clear budget constraint: to deliver several multi-sector clients in parallel without saturating my own days, I had to staff a small team fast - but without breaking the operating margin in the first year. The answer was the French alternance contract: hiring 2 developers and one project manager as work-study apprentices with a target of full autonomy in 3 months.

I built a recruiting funnel tailored to alternants - short technical exercise, posture interview, product-situation simulation - to filter motivation more than diploma. The first 2 weeks of each apprentice were dedicated to intensive pair-programming on the ACCENSEO codebases (the SaaS accounting platform and the internal tooling), with a written onboarding guide, progressive DEV then PROD access, and daily mentoring. From week 3 on, each apprentice owned a module end to end, with monthly code reviews and weekly objective check-ins. For the project manager, I duplicated the pattern on a real client engagement under mentorship, then released full autonomy once the project rituals were in place.

Within 3 months the 2 developers and the project manager were autonomous on their scope - able to frame a client request, deliver on their own, and run their own reviews. ACCENSEO's operating margin held because every apprentice took real product ownership from the first quarter, not a support role.

That experiment recalibrated the way I build a team: hire for the learning trajectory rather than for the current title, and accept that mentoring time is the highest-yield investment I can make at the early stage. It is the exact same logic I will replay in the next CTO scale-up role - a short hire-to-autonomy cycle is what separates a team that grows from a team that burns out.