Full Stack Developer @ ADEO - Freelance

ADEO Services is part of the ADEO Group, a major global player in home improvement retail. The group federates autonomous, interconnected businesses across many countries. Within the search & publication (CCDP) domain, the mission is to build and operate tools to publish and search products for Leroy Merlin business units worldwide.

I joined the Recommend & Push stream, which owns the group recommendation system: it combines data and algorithms (including ML), business knowledge (including the knowledge graph), and publication tooling so relevant products reach customers at the right time. The stream works end-to-end with developers, data scientists, data engineers, analysts, a product owner, and UX designers, in an agile setup focused on clean, stable, and scalable delivery.

Together with the team (pair and mob programming), my work is structured around four main areas:

Understanding the product and constraints

Map recommendation flows, API and BFF responsibilities, and how events and data move through Kafka and MongoDB. Align with data and product on use cases, non-functional requirements, and rollout constraints for multiple BUs.

Shipping features on the stack

Implement and evolve the recommendation product on the API, BFF, and front layers using Node.js, NestJS, and TypeScript, backed by MongoDB and Kafka. Contribute to integrations that keep publication and consumption of recommendations consistent and reliable.

Raising the bar on quality

Strengthen TDD and automated tests, improve CI/CD (including GitHub Actions), and clarify alerting and monitoring so regressions and incidents are visible early. Aim for maintainable code and predictable releases.

Operating as one team

Participate actively in ceremonies and technical discussions, share ownership of the product, and help spread good practices (software craftsmanship, code review, documentation) so the platform stays evolvable as traffic and features grow.

Beyond feature delivery on the recommendation surfaces, I contribute to how we observe and operate the system in production.

On the engineering side, this means tightening pipelines and checks in CI/CD, standardizing how we run and monitor services, and using tools such as Datadog (or equivalent) to track health, latency, and errors. The goal is to make the recommendation stack easier to reason about for on-call and day-to-day development.

On the product and data side, I collaborate with analysts and the data organization so that recommendation performance and usage can be measured and improved over time—similar in spirit to providing countries with clear indicators on the Decathlon app, but here focused on recommendation quality, coverage, and business outcomes across BUs.

When relevant, I also help improve internal or back-office experiences (configuration, publication workflows, or operational views) so teams can adjust and validate recommendation behavior without unnecessary friction.