Charles Gautier.

I graduated from Arts et Metiers, Angers campus, cohort 212. Angers is one of the school's historic places, installed for over two centuries in La Doutre. The school itself dates back to 1780, in a history deeply tied to industry, technical transmission and engineers trained to act in the real world.

I also carried rare student responsibilities: president of the student association in Angers, then national president of the Union des Eleves Arts et Metiers, the federation representing students across the campuses. It meant representing a collective voice, working with local teams, sitting in governance settings and speaking with the administration, alumni and experienced industrial profiles.

That period gave me a very concrete culture of management: budgets, trade-offs, representation, transmission, trust and the tension between tradition and long-term strategy. It still explains how I work today: a system does not hold only because it is intelligent. It holds because responsibilities are legible and because the humans carrying it can trust the frame.

The Arts et Metiers blouse is not decorative clothing. Historically, it participates in a simple idea: soften social markers, put students on the same level and remind everyone that what matters first is the person, their engagement and their place in the collective.

Each student then personalises their blouse through school life. Mine also tells my creative side: I painted it myself, inspired by an artist I liked, and it carries traces of that collective life, nicknames, family numbers, messages from friends, relatives and alumni.

It is a precise object for understanding how I look at systems: a shared frame strong enough to connect people, but open enough for each person to inscribe their own trajectory. This is what I now look for in AI architectures: structure without erasing the human.

After school, I carried HeartBoxGames, a mobile video-game studio. We developed a little under 150 concepts and prototypes with a demanding logic: move fast, test, observe, simplify and start again.

The hyper casual market teaches useful brutality. A player must understand in seconds, with very little attention available. If the gesture is unclear, if feedback comes too late, if the loop is not legible, the experience collapses.

That period became a school of crash-testing and tooling: spotting trends, building internal leverage, shortening cycles, making teams work, discarding what does not hold and shipping despite uncertainty.

It left a discipline that still structures my work: an idea is worth nothing alone. A system must run, be understood, be taken over, be improved and be governed.

From 2022, I started experimenting more intensely with AI, automation, marketing operating tools and the new technology layers already transforming how we build, sell, learn and operate.

The arrival of agentic systems shifted the question. We are no longer only speaking about a tool that helps write or automate a task. We are speaking about systems able to read, synthesise, act, parallelise, produce drafts, compare scenarios and make complex work far more legible.

But the more leverage increases, the more the human question matters: who decides, who validates, who understands, who resumes, who answers? My work is to build the harness around that power: context, tools, memory, permissions, confirmations, traces and escalations.

This is why the trajectory is coherent: collective work, management, production, prototyping, technology and governance. Agentic architecture is not a rupture in my path. It is the lever that makes this logic much more powerful.