09/10/2026 updated
100 % available
Senior / Lead ML Engineer — Personalization & Recommender Systems
AIX EN PEVELE, France
AIX EN PEVELE +50 km
Engineering Degree in Computer Science and Statistics — Polytech Lille, 2002LLMMLOpsGCPE-CommerceMachine LearningRecommender SystemsData ScienceApache SparkGitKubernetesMachine Learning OperationsTerraformDocker
Senior Data & Machine Learning Engineer with 20+ years of experience in software, data and AI, including 15+ years in Data Science and Machine Learning.
Strong expertise in large-scale recommender systems and personalization for e-commerce and retail: collaborative filtering, embeddings, product similarity and complementarity, ranking, cold-start and long-tail recommendation.
Several years of hands-on production experience on Google Cloud Platform, including BigQuery and Vertex AI, with Python, SQL, dbt, Airflow/Cloud Composer, Kafka, Docker and Kubernetes.
Strong ML Engineering and MLOps background: production ML pipelines, APIs and model serving, CI/CD, automated testing, monitoring, observability, data quality, cost optimization and scalable architectures.
Experience designing and evaluating A/B tests for recommendation algorithms, including randomized traffic allocation and session-consistent experiment assignment.
Technical leadership experience as former R&D Director and Senior ML/Data Engineer, working with Data Scientists, Data Engineers, software engineers, product teams and business stakeholders.
Currently building RecoKit, an e-commerce recommendation platform combining collaborative approaches, embeddings, knowledge graphs, vector search and LLM-based cold-start recommendation.
Daily use of Codex as an AI coding agent within a structured development workflow including Git, pre-commit hooks, mandatory unit tests, code review and GitHub Actions CI validation.
Python • SQL • Recommender Systems • Personalization • Ranking • Embeddings • GCP • BigQuery • Vertex AI • Spark/PySpark • Kafka • Kubernetes • Docker • MLOps • A/B Testing • GenAI • LLM • RAG • PostgreSQL/pgvector • Redis • ArangoDB
Experience
Professional experience10+ years
Freelancer since2017
Languages
EnglishBasic knowledgeFrenchNative speaker
Project history
Building an e-commerce recommendation engine focused on cold start, combining collaborative approaches, embeddings, knowledge graphs and LLMs. Designed a RAG architecture using Gemini File Search to leverage product documentation and enrich model context. End-to-end industrialization with Python pipelines and workers, APIs, asynchronous processing, vector storage, caching, monitoring and recovery mechanisms. Optimized AI processing costs through candidate pre-selection, caching and restricting LLM calls to cases where they provide additional value.
Teaching software engineering and data topics including programming, automation and testing, Git/CI, Shell, Airflow, UML, Excel/VBA and databases. Supervising students and explaining technical topics to audiences with varied backgrounds.
Designed, redesigned and industrialized large-scale e-commerce recommender systems deployed across multiple international Business Units. Developed recommendation and ranking approaches covering product complementarity, text/image similarity, embeddings, cold start and long-tail scenarios. Industrialized ML workloads on GCP using BigQuery, Vertex AI, Python, SQL, dbt, Airflow, Kafka, Docker and Kubernetes. Implemented technical and business monitoring covering data quality and completeness, logs, costs, catalog coverage, CTR, Add-to-Cart and conversion. Contributed to a major increase in the business impact of recommendations on Leroy Merlin websites between 2022 and 2024.