Expertise
What I can do
I am a fullstack developer. In practice: I design, build and ship web applications, and I stay with them once they are live. I have been doing this for 7 years on very high traffic sites (Groupe LaCentrale, Carrefour Voyages, Carrefour Services), from scoping the need through to production monitoring. Below are the six things people hire me for, each with something I have already delivered.
- Current role
- Fullstack AWS Developer, Groupe LaCentrale
- Experience
- 7 years in production, including 1 year as lead developer
- Location
- Paris area, France (on site or remote)
- Core stack
- TypeScript, React, Next.js, Node.js, AWS
1.2 B
requests a month on LaCentrale's search API
16 ms
average response time on that same API
+40 pts
mobile Core Web Vitals gained at Carrefour Services
3 M
visitors on the SEO pages I generate at LaCentrale
Building a web app from scratch
You have an idea, a business need or a project to take over. I handle the whole thing (scoping, design, development, launch), on my own or inside a team.
- Fast, accessible web interfaces (React, Next.js, TypeScript)
- Back end, APIs and databases (Node.js, PostgreSQL, NoSQL)
- Deployment, monitoring and follow-up work after launch
Already delivered
At Groupe LaCentrale I carried the SEO CMS platform from scoping and architecture through to delivery and monitoring, moving forward with input from the product and SEO teams. It now generates over 8,000 URLs a month.
Making a site fast and visible on Google
A slow or poorly indexed site loses visitors before it has shown them anything. I measure, I fix, and I prove the gain with numbers.
- Core Web Vitals, Lighthouse, load times, accessibility (WCAG, RGAA)
- Technical SEO: structured data, rich snippets, internal linking, programmatic SEO
- Reporting dashboards so the gains stay visible over time (Looker Studio)
Already delivered
+40 points of mobile Core Web Vitals at Carrefour Services, with the last-minute page 30% faster. At LaCentrale, programmatic SEO reached the Google top 3.
Putting AI into a product, for real
Not a demo that impresses in a meeting. AI features that run in production, whose quality is measured and whose cost is under control.
- Natural language search instead of the usual filter panels
- Scoring and re-ranking to make results genuinely relevant
- LLM integration: AWS Bedrock, function calling, structured outputs, Langfuse
Already delivered
At LaCentrale I designed a natural language AI search engine and an AI scoring layer that makes results more relevant. Certified "Developing Generative AI Applications on AWS" in 2025.
Handling the load: architecture and AWS cloud
Designing systems that stay fast and available as traffic grows, without the cloud bill growing with it.
- AWS and serverless architecture (CDK, CloudFormation, Docker)
- Fast REST APIs, caching and response time optimisation
- Monitoring, alerting and high availability
Already delivered
LaCentrale's search API takes 1.2 billion requests a month at a 16 ms average response time. At Carrefour Voyages I rolled out the monitoring and alerting that keeps the sites available.
Structuring a team and levelling it up
When a team grows, the process usually breaks before the code does. I put in the rules that let a team ship quickly without breaking production.
- Managing developers and ramping them up (React, TypeScript)
- Definition of done, test plans, unit and E2E tests, code reviews
- Code quality and security, CI/CD (GitHub, GitLab, Sonar, Wiz)
Already delivered
Lead developer at Carrefour Services for a year, in a team of 3 developers. I ran the React ramp-up and put in the development process the team needed to scale.
Deciding with data rather than instinct
Understanding what visitors actually do, testing changes before rolling them out, and keeping only the ones that work.
- User behaviour analysis (Contentsquare, PostHog, Hotjar, GTM)
- End to end A/B testing (Kameleoon, AB Tasty, Beyable)
- Dashboards and reporting (Looker Studio, Tableau)
Already delivered
Contentsquare certified (2023, then Find & Fix in 2024) and Kameleoon certified (2024). Analyses and A/B tests run on sites with millions of visitors.
How I work
- 01
Scoping
I start from the problem, not the solution. Goal, users, technical constraints, existing code to take over. What is out of scope is written down as plainly as what is in.
- 02
Design
Architecture, data model, user journeys and technical choices settled before the first line of code. You know what is going to be built, and why.
- 03
Development
Short iterations, a preview environment online, automated tests and code reviews. Every step is visible and testable, not just the last one.
- 04
Going live
Automated deployment, monitoring and alerting in place, documentation up to date. The project runs, and it runs without depending on me.
Stack & tools
Languages & frameworks
TypeScript, JavaScript, Node.js, React, Next.js, PHP
Data
PostgreSQL, MongoDB, SQL, NoSQL, Prisma
Cloud & infrastructure
AWS, GCP, CloudFlare, OVH, Serverless, CDK, CloudFormation, Docker
Architecture & patterns
Microfrontend, Clean Architecture, DDD, Hexagonal, MVC, REST APIs, TDD
Quality & CI/CD
GitHub, GitLab, Bitbucket, Jest, Sonar, Wiz
Analytics & A/B testing
Contentsquare, PostHog, Hotjar, GTM, OneTrust, Kameleoon, AB Tasty, Beyable
AI
AWS Bedrock, Langfuse, function calling, structured outputs, prompt engineering
Design & project management
Figma, Adobe XD, Jira, Trello, Agile, SCRUM
