{ whoami }
I build the platforms other software stands on — Kubernetes-native, event-driven, multi-tenant systems — and I build agents on top of them. Platform engineer by trade, AI enthusiast.
Based in Grand-Est, France, open to hybrid roles in France, Belgium, and Luxembourg, or remote elsewhere.
Focus
Platform · AI
Stack
Go · TypeScript · Python · Kubernetes
Exp
4+ Years
{ the story }
Graduated the bachelor programme at La Capsule, repeatedly rated the best coding bootcamp in Europe. The full stack, end to end: TypeScript, React, React Native, Node, MongoDB.
Taught web & mobile development to the next cohorts for a year and a half in Lille and remotely. In parallel, built booking platforms for local businesses — nci coiffure and OsteoBook — both still live and serving clients today.
A Heroku-like, event-driven platform on an on-prem Kubernetes cluster: students deploy apps from a catalog into strictly isolated environments. Custom operators, GitOps reconciliation, gRPC + RabbitMQ microservices, and a bespoke scale-to-zero autoscaler built for thousands of idle environments.
3000+
self-healing environments
2
custom operators & CRDs
2+
years in prod
Passed the CPSA-F — Certified Professional for Software Architecture, Foundation level — while building Kapps'UL. Design decisions backed by method, not just instinct.
All 15 CNCF certifications plus LFCS in six months — Golden Kubestronaut #479 worldwide, #20 in France. Shipped VME & Uniq for the university, started Mantoa, and went deep on Go internals with Ardan Labs.
Built platform
2024–26Université de Lorraine · PLEIADES
A Heroku-like platform on an on-prem Kubernetes cluster: students deploy applications from a catalog into strictly isolated, quota-enforced environments. Event-driven microservices, custom operators extending the Kubernetes API, GitOps reconciliation, and a bespoke scale-to-zero autoscaler keeping thousands of idle environments cheap.
3000+
isolated tenants
2y+
in production
Université de Lorraine
2026Campus software observability: which software lives on which machine, what teachers actually use, what can be reclaimed — feeding automated park maintenance through Puppet automations.
Université de Lorraine
2026A unique naming and tracking system for the campus machine park — one canonical identity per computer, from enrollment to retirement.
Freelance — barber shop
2023Booking platform for a barber shop — online appointments, services, gallery. In production and booking real clients since launch.
Freelance — osteopath
2024Booking platform for an osteopathy practice with the same goal: zero phone tag, clients can book their own appointments. Still live, still used daily.
Apps in production
Open source at v0
2026 — Work In ProgressAgentic learning coach: feed it a book, course or technical document — it builds the roadmap, runs your daily sessions, tests you and schedules reviews until mastery. Document ingestion, knowledge extraction, LLM orchestration, graph + vector stores.
Agentic AI & Cloud Platform
Side quests
Platform
2026How to organize multiple watchers of Kubernetes cluster resources, potentially triggered by more than 3,000 users at once, for a real-time UI without overwhelming the API.
Blog Posts
See my work
Read my blog
{ the cncf journey }
#479
worldwide
#20
in France
15 + 1
CNCF certs · LFCS
6
months total
{ depth }
Certified Professional for Software Architecture — Foundation
I passed a software architecture certification to validate formally what Kapps'UL demanded daily: trade-off analysis, architecture documentation, quality goals and design under constraints — decisions backed by method, not instinct.
Go, at the metal
The more AI assistance we get, the deeper our fundamentals must go. That's why I trained with Ardan Labs, one of the most renowned providers of Go training, on mechanical sympathy and Go internals — the language the cloud is written in.
v0 — work in progress{ building now }
Most people don't fail to learn because they lack information — they fail because they don't have a system. Mantoa turns any book, course or technical document into a complete learning program: a structured roadmap, guided daily sessions, understanding tests, perfectly-timed reviews, and explanations until mastery.
You provide the effort. The app provides the method. Document ingestion, knowledge extraction, LLM orchestration, graph & vector stores — evolved from ULearn, open source at launch.
{ convictions }
Opinions on AI-assisted engineering
I’m fully in favor of AI assistance—with caveats. I use it to move faster across the parts of engineering where speed is useful
I deeply believe in open source. While we wait for self-hosted models to become powerful and fast enough for everyday work
We too often use AI to optimize shipping speed instead of deepening our understanding. Long-running
Available for new challenges
Platform engineering, cloud-native products, agentic systems — if it needs to scale and stay understandable, I'm in.
