Abdeslam S.

Project Objective: Implementation of a data product within a Data Mesh architecture to verify the compliance of core banking and mortgage systems for AML/CFT (Anti-Money Laundering/Combating the Financing of Terrorism) requirements.

Implementation of dbt pipelines to sample and compare customer data across multiple systems, enabling cross-domain validation. Implemented BigQuery as the central ingestion layer and data sink, structured according to a Medallion architecture to load samples, historize comparison results, and produce gap analyses. Industrialized workflow execution in containers orchestrated by Kubernetes to optimize pipeline performance and resilience.

Project Objective: Implementation of a data product within a Data Mesh architecture to harmonize the extraction and processing of transactions, in response to TRACFIN (Financial Intelligence Processing and Action Unit) requirements for AML/CFT and the fight against financial crime. Implementation of a hybrid SharePoint-BigQuery pipeline automating near real-time ingestion of search data for the Data Product. Development of dbt pipelines to transform and model transactional data, meeting business requirements for fraud detection and compliance. Automation of the generation of regulatory TRACFIN investigation reports, based on analyses produced by dbt pipelines. Reduction of TRACFIN investigation processing time from 72 hours to less than 15 minutes, improving regulatory response capability.

Technologies Used: DBT, GCP (BigQuery, Cloud Storage, Cloud Logging), Kubernetes (IaC Helm/Kubernetes manifests), Python (SqlAlchemy, UV), GitLab CI/CD, Scrum.

Project Objective: Migration of technical and commercial data (network configurations, usage data, repositories) from Orange France customer equipment (set-top boxes, Livebox) to Google Cloud Platform (GCP). Design and implementation of the cloud architecture, including a study of cloud services aligned with architectural requirements. Creation of multiple data processing pipelines to meet various business use cases. Hybridization of the OnPrem/Cloud solution, allowing consumption of OnPrem services (REST API) and industrialization of querying processes. Industrialization of the solution through the separation of infrastructure/business code, ensuring robustness, scalability, and maintainability. Implementation of multi-dimensional monitoring indicators: data volume, pipeline performance, processing time, and availability.

Technologies Used: GCP (BigQuery, DataFlow, Cloud Function, Scheduler, Cloud Storage, Cloud Logging, Cloud Monitoring), Terraform, Python, GitLab CI/CD, Scrum.

Project Objective: Implementation of a geolocation platform for judicial requisitions to locate mobile users on 3G, 4G, and 5G networks. Implementation of a Kafka producer ensuring real-time ingestion of HDFS data to Kafka. Implementation of a Kafka consumer for real-time injection into ScyllaDB using

Project Objective: Proof of concept (POC) of a solution for real-time monitoring of ATM availability. Participation in the setup and execution of a lambda architecture. Implementation of a data streaming pipeline for data ingestion, processing, and storage using Python and Pyspark Structured Streaming. Pushing data to various sinks (NoSQL, relational database).

Technologies Used: Spark Structured Streaming, PySpark, Kafka, PostgreSQL, MongoDB, Elasticsearch, CentOS/Rocky.