The path of a transition manager turned AI Transformation Leader

A qualified engineer from the École des Mines de Nantes (2001), specialising in Information Technology Organisation and Management (OMTI), and holder of a law degree from the Université Paris-Sorbonne. This dual training (systems engineer and lawyer) was not a career plan, it was curiosity. Years later, it became the angle that lets me move from code to strategy without changing language. But before the degree, I had already been a developer for eight years.

Before the Mines: eight years of code

I started coding in 1993. Not as part of any training: out of curiosity and necessity. Between 1993 and 2001, before and during my engineering studies, I developed several software projects that trained me far more deeply than any course could.

• 1993-1995: Development of a student records management system. My first standalone software project, my first experience of a complete cycle (specification, design, development, deployment, maintenance).
• 1995-1998: An algorithmic image-processing module for national defence. My first industrial application of algorithms to constrained data streams.
• 1997-2000: Design of a genetic algorithm, in operations research, for a project supporting people with disabilities through volunteers. That is where I learned to formalise a business problem into constraints and to measure the quality of a solution.
• 1998-2001: Founding of Ajileso, an agile software development company. My first experiences of tech project management and team development.
• 1999-2001: The Maisonnez.fr project, handling large-scale data (Big Data before the term existed).

Eight years of practice before the engineering degree. The École des Mines formalised what practice had built. That is why today, when I talk to developers deploying AI agents, and when I talk to business leaders wondering about the impact of AI on their business model, I do not feel I am switching registers.

Consulting, transition, general management

My early professional years were marked by consulting. A consultant at Michael Page, an independent trainer, certified in Process Communication Management. Then, e-Business Consultant at Bureau Veritas (2001-2003), where I contributed, alongside Guillaume Klossa, to a submission under the aegis of the United Nations Economic Commission for Europe (UNECE, Geneva) on the regulation of e-marketplaces by a trusted third party and the prevention of monopsony positions on those platforms. At 25, it was this experience that taught me to articulate the rigour of an international normative framework with the operational reality of a business.

Very soon afterwards, I moved from pure consulting to operational management, taking the helm of companies in transition, that is, at critical moments where one must steer strategy, teams, finance and transformation all at once.

Fifteen years, seven sectors, one method

Transition management became my main discipline. Over fifteen years, I have led organisations in radically different worlds:

• Recruitment and HR consulting: heading the RhFed firm from 2005, more than 150 hires of engineer, technician and sales profiles.
• Training and research: the European Technological Institute for Music Professions (ITEMM, 2013-2015), the Sorbonne Foundation (2015), a training centre for living-history professions housed in the former premises of a Renault dealership in Chartres.
• Real estate and services: transformation projects for the Saint-Malo market (2011), structuring of service-sector activities.
• Events and business services: running events organisations, conferences, summer schools (notably 2FPCO).
• Specialist retail: a post-Covid pivot of a holiday-lodge business towards Nordic baths and saunas in the Sarthe (2021), a full-scale example of an operational pivot under constraint.

This diversity has a logic. In every engagement, the apparent subject is different: music, university, real estate, heritage, hospitality, recruitment. But the underlying craft is the same: taking the controls of an organisation at the moment it must change, and making sure that change succeeds. Diagnose, scope, execute, measure, hand over. These five steps are the foundation of everything I do today.

The pivot to AI

Since 2023, this foundation has applied to transformation through artificial intelligence. It is the subject that demands the most of that discipline. AI tools change every six weeks, the promises are enormous, and so are the disappointments. What distinguishes AI transformations that succeed from those that end up in the “POC graveyard” is not the technology. It is the method.

Today I devote most of my time to designing and operating orchestrated AI agent systems, run from an entirely sovereign infrastructure. Alongside my engagements, I operate several SaaS platforms in production (Junyr ERP, Junyr Mail, Maisonnez.fr, Timbry, Sepialy), all of them testing grounds for the AI agent architectures and sovereign infrastructure I deploy with my clients.

This stance (pragmatic, vendor-independent, compatible with the confidentiality requirements of French SMEs) is carried by the SAS Croissance et Transitions, a company I created in March 2026 to pursue this specialisation full-time.

References

• Listed AI trainer with the Fédération des Objets Média (FOM): the professional federation of the media-object and promotional-object sector. Officially listed as a trainer on artificial intelligence topics.
• UNECE contribution 2001-2003: United Nations Economic Commission for Europe (Geneva), on the regulation of e-marketplaces by a trusted third party and the prevention of monopsony positions, alongside Guillaume Klossa, in my role at Bureau Veritas. Public document on unece.org →
• Head of the RhFed firm (~2005): a recruitment firm, 150+ hires of engineers, technicians and sales staff. Cited in the trade press (Actionco, 2005 and 2007).
• Managing director of ITEMM (European Technological Institute for Music Professions, 2013-2015). Listed on the institutional website.

A discipline that is cultivated

Alongside consulting, I run. Regularly, on road and trail. I do not make it a professional argument, but it is not a detail: long-distance running teaches what AI transformation demands: consistency, acceptance of time, and the ability to hold a course when the subject becomes uncomfortable.

My commitments

• No deployment without written scoping.
• No scoping without an identified, committed end user.
• No scaling up without documented measurement of the benefits of the previous phase.

Engagements that depart from these three principles are not supported.