Taha Yassine Benbrahim

Design and implementation of a modern Data Platform architecture in Medallion architecture (Bronze -> Silver -> Gold) on Spark and dbt-spark, with a strategic focus on Data Quality and the reliability of actuarial indicators

Main achievements:

* Data Architecture in Medallion: ingestion of raw data (Bronze), cleaning and standardization (Silver), then valuation (Gold) for actuarial, reporting and regulatory needs.

* Large-scale Data Quality: implementation of more than 120 DQ rules (completeness, uniqueness, accuracy, consistency) via an automated framework (dbt macros + YAML generated in Python).

* UC Entreprises Actuarial Projects (2019–2024): development of reconciliation pipelines between actual and estimated premiums (DIM, PE, etc.) on key categories (60, 61, 63, 66) and several guarantees (Death, Disability, Illness…).

* Orchestration and Industrialization: automation of flows via Airflow (32+ DAGs) and deployment of models on a Spark cluster integrated with MinIO and Dremio.

* Optimization & Business Support: data reliability for actuaries, control of financial discrepancies, implementation of automated monitoring reports.

Integrated into the Risk department, within the BDSI Datalake team, with the responsibility of designing, industrializing and maintaining Big Data pipelines for regulatory reporting and compliance.

Main achievements:

Ingestion & Storage: integration of data from SAB, storage in parquet format and creation of associated tables under Hive.

Data Factory: feeding and transforming data for reporting needs (Cognos, Oracle).

Industrialization of processing: implementation and automation of ingestion pipelines, ksh scripts for deployments via Jenkins.

Support & Skills Transfer: supporting Overseas teams in the transfer of the application scope.

Optimization & Refactoring: improving the performance and quality of existing code, resolving incidents and anomalies.

Inter-team exchanges: generation and transfer of CSV/PDF files for integration by other entities via CFT.

CI/CD & Governance: validation of merge requests, estimation of changes and contribution to the technical roadmap.

Management and industrialization of strategic projects around risk management and climate transition, in a hybrid GCP + Azure environment.

Main achievements:

📌 Project 1: Continuous Risk Monitoring (SCR)

* Integration of external data (Factiva Dow Jones, INSEE) to enrich the counterparty repository.

* Development of real-time pipelines with Event Hub + Spark/Databricks for ingestion and transformation of flows.

* Creation of models and KPIs calculated in PySpark, stored and used via Azure SQL.

* Indexing and making data available via Azure Search for Front-End use.

* Participation in the industrialization of ML models for anomaly detection.

* Code review, validation of merge requests and maintenance in operational conditions.

📌 Project 2: Physical Climate Risk (RCP)

* Design and implementation of the technical architecture of the RCP project on GCP.

* Development and orchestration of processing via Airflow + PySpark.

* Implementation of CI/CD pipelines (Jenkins, Terraform, Ansible) and automation of flows (CTRL-M for network transfers → GCP buckets).

* Deployment in production and monitoring of acceptance testing.

* Design of star schemas for BI and regulatory reporting needs.