Philippe Paulos

Scaled a B2B data platform within Sodexo's Digital & AI division, enabling clients to monitor on site performance through Power BI dashboards. The product evolved from a single client proof of concept into a configuration driven framework now deployed across ~20 clients and growing. Led requirements scoping and technical discovery with business users across a base of ~20 client deployments, framing each new data flow with source-system teams before any pipeline was built. Built and operate ETL pipelines integrating heterogeneous business sources (point-of-sale, purchasing, hygiene audits, customer satisfaction) via REST/SOAP APIs, Microsoft GraphAPI, SharePoint, Dremio and file-based exchanges Owned the dbt platform end to end: repo, Slim CI, Databricks integration, shifting heavy Power BI SQL logic into governed, incremental dbt models, improving dashboard performance and maintainability Designed and shipped to production an end-to-end GenAI solution automating meal categorization, replacing a fully manual monthly process and saving ~2 days/month. Automated data quality and freshness controls, job monitoring and alerting (Pandera, Grafana) Coordinated the data engineering team: code reviews, owner of coding standards and best practices Stack: Python, Databricks, dbt, Azure, PySpark, Polars, PowerBI, Dremio, Azure OpenAI, Azure Document Intelligence, CI/CD, GraphAPI, REST/SOAP APIs, Grafana

Led the data engineering efforts for a scheduling enhancement project for SPSO vessels, used by Total subsidiaries in Nigeria, Congo, Denmark, and Angola. The system supports operations producing up to 25 million dollars of barrels of oil per day per site, making scheduling and resource planning critical. Designed and built data integration pipelines from multiple sources, improving the accuracy of scheduling and overall operational efficiency. Led the development of data starter kits to accelerate MVP delivery (Azure, Databricks, AWS, Spark). Stack: Microsoft Azure · AWS · Terraform · Python · API REST · Spark . Databricks · AWS Lambda · Azure Functions · DBT . SQL · docker · Oracle Primavera · Module SAP PM

Unification of ERP supplier databases across Europe, the USA, and Asia at Air Liquide, creating a centralized data ecosystem to improve decision-making, purchasing efficiency, and supplier management. This consolidation is estimated to generate 5–10% in annual procurement savings, primarily through improved supplier negotiations, elimination of duplicate contracts, and reduced administrative overhead. Automated the integration of multiple ERP data sources into a single, consistent database shared to the whole company (snowflake). Developed reconciliation scripts to standardize contracts and supplier address data, supporting a complete and accurate migration of supplier records. Built analysis scripts to identify discrepancies, overlaps, and opportunities for contract consolidation. Implemented CI/CD pipelines to ensure reliable deployment of Python data processing scripts. Stack: Apache Airflow · Python · GCP · Talend · GitLab · docker · Microsoft Power BI · PostgreSQL · DBT . SQL · Snowflake · Oracle ERP · SAP ERP

Designed and implemented a GDPR compliance module for AXA’s data lake, covering over 100 data sources. The solution enabled consistent migration and transformation of sensitive data using a fully configurable, YAML-driven approach, significantly reducing manual engineering work. I led the full design, development, and orchestration of the module, ensuring compliance across diverse ingestion pipelines. Migrated contract and finance data warehouses to the Azure platform. Delivered business-specific data views by preparing, calculating, and exposing data across multiple analytical dimensions. Stack: Microsoft Azure · Apache Spark · Python · Databricks · Azure Data Factory · Azure DevOps · DevOps · Microsoft Power BI · Règlement général sur la protection des données (RGPD)

Led the migration of banking transaction applications to an on-premise data lake within the R&D team, enabling premium control and real-time analytics (ePortal) for major clients including IKEA, Leclerc, and Sainsbury's. Developed and maintained over 20 ingestion, processing, and enrichment jobs for real-time transaction data. Designed and implemented CI/CD pipelines to streamline deployment and reduce operational friction. Led a team of five data engineers and two QA engineers, ensuring delivery quality and team efficiency. Facilitated workshops with development and testing teams to improve code quality, job maintainability, and delivery velocity. Collaborated closely with SysOps teams to enhance monitoring and troubleshooting tools for production workflows. Participated in the design of new data platform components in coordination with the architecture team. Stack: Scala · Kafka · Python · Apache Spark Streaming · HDFS · ElasticSearch · Hadoop · Scrum · GitLab · Oracle Database · SQL · Hive · NoSQL · HBase

Within the Network and Analytics IT department, integration of data from mobile probes to provide indicators for improved network supervision. Development: Design, implementation, and optimization of data ingestion processes from mobile network probes (incoming rate of 10,000 messages/s). Exposure of KPIs to customer services and supervisions (REST API). Design of the architecture for new projects. Sysops: Administration of the Hadoop platform. Cluster upgrade, addition of Datanodes, version updates, high availability of services. Automations via Ansible (adding Datanodes, user management), continuous deployment. Stack: Scala · Apache Spark Streaming · Cloudera · Administration système · Spring Boot · API REST

Stack: HBase · Apache Spark · Scala · Spring Boot · API REST · NoSQL