Othmane Benbella

Macif is one of France's largest mutual insurance groups, undertaking a strategic data platform modernization with a large-scale migration from on-premise ETL system Informatica Powercenter to the Cloud on BigQuery, using dbt for the data transformation worlflows.

I lead the dbt governance and industrialization program supporting this transformation, working across multiple analytics teams and dbt Cloud projects.

- Led enterprise-wide dbt governance program, establishing reference standards adopted across all analytics teams

- Architected a shared GitLab CI/CD pipeline with automated quality gates (SQLFluff linting, dbt Project Evaluator, tests, documentation coverage), reducing onboarding time for new dbt projects to under one day

- Built a production-grade reusable dbt project template, now adopted organization-wide as the baseline for all new analytics initiatives

- Authored Macif's BigQuery performance and cost optimization guidelines (partitioning, clustering, query patterns, incremental strategies)

- Designed a FinOps tagging and cost-monitoring strategy for BigQuery, providing detailed cost visibility for budgeting decisions

- Developed an internal dbt package tracking dbt run execution and load monitoring across environments

- Designed and delivered a three-tier dbt training curriculum through hands-on workshops, covering analytics engineers, data analysts and functional stakeholders

- Facilitated weekly code review clinics and office hours, promoting best practices and increasing team autonomy

- Codified Macif's dbt development and documentation standards as automated CI/CD quality gates (linting, tests, doc coverage, governance audits)

Stack:dbt Cloud, BigQuery, GitLab CI, Python, SQL, dbt Project Evaluator, SQLFluff

BforBank is a French online bank operating in a regulated environment with strict requirements for data traceability and reporting accuracy.

I led the migration of legacy Informatica PowerCenter workflows into a structured dbt project on BigQuery, ensuring functional parity with existing production logic and full reconciliation against legacy outputs.

- Translated complex Informatica mappings into a layered dbt architecture (staging, intermediate, marts), preserving business logic precisely including joins, filters, routers, update/insert strategies and column ordering

- Reconstructed dependency graphs from PowerCenter connectors and aligned them with dbt DAGs, ensuring orchestration semantics were faithfully reproduced

- Conducted systematic data validation tests confirming 100% parity between cloned Oracle source snapshots and the produced BigQuery tables

- Designed and industrialized data products generating XML legal and financial reports, replicating legacy PowerCenter logic in Python for downstream regulatory submission

- Implemented complex incremental model patterns, including handling of late-arriving data and update/delete semantics aligned with banking source systems

- Collaborated closely with business stakeholders to clarify edge cases, confirm orchestration intent and iterate on validation findings, ensuring accurate reproduction of the source workflows

- Documented mappings, lineage and validation outcomes to support audit and reconciliation processes

Stack:dbt Cloud, BigQuery, Oracle SQL, Python, Informatica PowerCenter (source), Git, GitLab CI

Scopely is one of the world's largest mobile game publishers, operating multiple game studios with high-volume event analytics pipelines on BigQuery. I joined to design and deliver an AI-accelerated migration framework converting legacy Airflow SQL workflows into clean, dbt-compliant equivalents, while supporting live production analytics during the transition.

- Designed and implemented an AI-accelerated migration pipeline converting legacy Airflow SQL workflows for game analytics, including deprecated source mappings and complex joins, into dbt-compliant equivalents on BigQuery

- Deployed containerized MCP (Model Context Protocol) servers for dbt, BigQuery, FileSystem and Markdown, orchestrated via Docker Compose for local development environments

- Integrated MCP servers into VS Code through stdio bindings for local engineering use, and exposed them over SSE using FastAPI for external service access

- Built a Python-based dbt and Looker lineage extraction toolkit, capturing model-level and column-level lineage across dbt projects, Looker dashboards and legacy SQL workflows

- Exported lineage results to Excel for tabular review and produced interactive visual maps, helping anticipate Looker PDTs requiring migration, flag dashboards at risk from refactors, and plan safe changes

- Debugged and optimized BigQuery SQL queries under production deadlines, supporting live analytics pipelines for Scopely's game studios during the migration window

- Wrote dbt models for the migrated event datasets, ensuring uninterrupted reporting and analytics continuity throughout the transition

- Combined senior dbt expertise with AI tooling to compress migration timelines while preserving correctness and lineage traceability

Stack: dbt, BigQuery, Python, FastAPI, Docker, Docker Compose, MCP servers, VSCode, GitHub, Looker, Make, SQL, Airflow