Philippe Paulos

For the past year, I have been working as a Tech Lead Data Engineer within Sodexo's Digital & AI division, on a B2B data product that allows group clients to track the performance of Sodexo sites deployed at their locations.

Specifically, the tool exposes a wide range of KPIs: hygiene audits, food waste, transactions (admissions, tickets), capture rates, supplier purchases, etc. All of this is presented in PowerBI dashboards for clients.

I lead a small team of three people (2 data analysts and 1 data engineer) and work 80% hands-on. My role covers the entire data product chain: ingestion of data from multiple sources, transformation, modeling, and exposure to the reporting layer.

I am also the technical point of contact for the Product Owner and business sponsors, I lead design meetings and technical breakdowns, I manage technical debt, and I support the skill development of junior profiles.

The entire chain is tested and automated: CI/CD on Azure DevOps with environment promotion, linting and pre-commit hooks, systematic code reviews. On the observability side, data quality and freshness controls are implemented per client, with schema validation and processing status reporting to supervision. Anomalies are thus detected before they appear in client dashboards.

Technically, the environment is primarily Azure: Databricks and Spark for large-scale processing, Azure Data Factory for orchestration, Azure Data Lake for storage, and Python (Pandas, Polars) for data manipulation.

Contribution to the major project of redesigning FPSO ship planning for Total subsidiaries (Nigeria, Congo, Denmark, Angola).

The system handles operations producing up to **$25 million barrels of oil per day per site**, making planning and resource allocation essential.

Unification of ERP supplier databases in Europe, the United States, and Asia to create a centralized data ecosystem for improved decision-making, procurement efficiency (**5 to 10% annual savings**), and supplier management.