Ernest Popovici

At Effy, I participated in modernizing the analytical layer by introducing a Semantic Layer based on MetricFlow on top of an existing dbt project. I migrated and refactored dbt models, defined semantic entities, and standardized business metrics used in dashboards. This reduced inter-team discrepancies, strengthened trust in data, and prepared the ground for true self-service analytics. I worked directly with data and business teams to align metric definitions with business reality.

At EDF Renouvelables, I participated in a long-term project to build a complete Data Lab on AWS. The objective was to centralize data and analytics for several internal teams. I first configured an initial data integration platform, then led the migration to a more advanced solution. This involved refactoring ingestion pipelines, adapting data models, and close integration with AWS services like S3, Glue, Athena, CDK, or SageMaker. I also wrote documentation, led training sessions for technical and non-technical profiles, and served as an English-speaking point of contact with external partners.

At Datanumia, I worked on industrializing the analytical platform. The existing solution relied on numerous Python scripts orchestrated in AWS Glue, which were difficult to maintain and evolve. I introduced dbt as the central transformation framework on top of Redshift, redefined the modeling layer, added tests and documentation, and contributed to configuring the Redshift cluster as well as orchestration via Lambdas. The new architecture significantly improved stability, transparency, and development speed for future enhancements.