Taha Youssef

- Cloud Architecture: Design and steering of the target cloud architecture (GCP) for Data Warehouse migration: definition of the multi-environment model (5 stages), standardization of Terraform patterns (network, security, compute, data), and scoping of structuring technological choices.

- Resilience & DR: Design and implementation of the platform's disaster recovery strategy: definition of RPO/RTO and SLA objectives per service, automated backup (PostgreSQL/Airbyte to GCS) with RPO compromise alerting, and controlled upgrade procedures (Helm versioning, Blue/Green). Formalized DR in runbooks and validated by recovery exercise: RTO measured at 10 min.

- Security & Multi-tenancy: Design of a multi-subsidiary isolation model (project-per-entity topology) with per-user data access control: row access policies, authorized views, and authorized datasets designed for scalable CI/CD deployment. In production for a first subsidiary, generalizable to the entire scope.

- Standards & CI/CD: Definition of CI/CD architecture and deployment standards (Bitbucket Pipelines) covering dbt, Terraform, etc..

- Microservices Architecture: Design of a microservices architecture (FastAPI / Cloud Run) for the Report Automation project: technical arbitration, processing parallelization, observability (logs, alerting, retries).

- Kubernetes Platform: Steering of the shared Kubernetes cluster (GKE): deployment, operation, and migration of Cloud Run applications to GKE (ArgoCD, Kestra, Prometheus).

- Governance & FinOps: Implementation of Data Lineage (OpenLineage), data governance (IAM, audit logs, quality), and FinOps optimization with BigQuery cost reduction from €13k to €7.5k.

- Business Impact: Transverse collaboration to provide reliable data for anti-fraud, credit scoring, debt collection, and customer segmentation.

- Regulatory Scope: Data referent for regulatory reporting (Banque de France, DSP2, OSMP, BCE), with functional and technical responsibility for fraud indicators covering over 3 million consolidated indicators on GCP/BigQuery.

- Migration & Optimization: Steering the strategic Spark -> BigQuery Engine migration, reducing overall computation time from 45 min to 10 min on critical chains.

- Secure Pipelines: Design of secure pipelines (ingestion, quality controls, encryption, traceability) ensuring regulatory compliance and auditability.

- Orchestration: Architecture and orchestration of complex batch processing via Apache Airflow and Dataproc (SLA, incident recovery, dependencies).

- Business Tooling: Development of a secure internal web application (CloudSQL + BigQuery) for entering and validating regulatory indicators.

- Coordination: Interface between data, fraud/payment business, and compliance teams, translating regulatory needs into robust data solutions.

- Development of a real-time platform for vehicle telemetry: Kafka ingestion → Spark Streaming → Elasticsearch.

- Calculation of performance indicators and feeding of Kibana dashboards for predictive diagnostics.

- Implementation of data observability and processing alerts.