Rida Layadi

Within the Data Collection team, I develop robust pipelines for collecting and ensuring the quality of internal and external data. My main responsibilities:

• Data collection from internal OVH and external sources (Google, Facebook, LinkedIn)

• Development of distributed pipelines with Apache Spark, orchestration via Airflow

• Implementation of data quality controls (validation, filtering, completeness)

• Containerization of jobs with Docker and deployment via Kubernetes

• Use of Dremio for data exploration and provisioning

• Storage on OVHcloud Object Storage (S3-compatible solution, independent of AWS)

This role is part of a demanding sovereign cloud environment, where data reliability, process automation, and solution scalability are key challenges.

Project: Anti Money Laundering Project

Methodology: Agile Scrum

Context:

As part of the ABS (AML Behaviour Spotting) project, the FCC (Financial Crime Compliance) department needs a tool to identify atypical behaviors from a transaction source. In summary, this project allows for:

• Contributing to the improvement of LAB-FT (Anti-Money Laundering and Counter-Terrorist Financing) activities.

• Consolidation of customer-level operations.

• Identification of new trends that may hide money laundering risks.

• Taking preventive actions to target abnormal behaviors.

Missions and tasks performed:

• Development of Spark scenarios: Creation and optimization of Spark scenarios for detecting atypical behaviors, particularly related to crypto transactions and terrorist financing for Société Générale, Boursorama, and Crédit du Nord banks.

• Analysis and Design: Analysis of business needs and specifications, technical solution design, development, testing, and evolutionary and corrective maintenance of functionalities.

• Data Collection and Processing: Collection of data from various sources into a Data Lake, processing data using Spark, and generating signals in a PostgreSQL database.

• Signal Generation: Development of processes for detecting atypical behaviors, enabling the generation of signals that are stored and analyzed in PostgreSQL. Signals are calculated from bank transactions using various algorithms such as full match, regex, and word-in-word, based on terms retrieved via a user interface that the business changes regularly.

Technical Environment: Agile Scrum, Scala, Spark, Ansible, AWS, PostgreSQL, Jenkins, Git, GitHub, IntelliJ, Maven, Jira, Hue, Confluence, Docker, Kubernetes, NestJS, ReactJS.

Client: STET, La Défense

Project: INSTANT PAYMENT Project

Methodology: Agile Scrum

Context:

Created by six French banks, STET handles the clearing of retail payment methods (transfers, direct debits, check images, card transactions, etc.) between all banks in France. It also facilitates SEPA transfers nationally and in Europe. The "Instant Payment" project involves designing a real-time payment system intended to process 20,000 transactions per second with a latency of around one second and very high availability (99.99%).

Missions and tasks performed:

• Participation in the development of business features for Instant Payment interbank transfers (real-time European bank transfers) and SEPA Direct Debit (SDD).

• Analysis of specifications, design, development, testing, and evolutionary and corrective maintenance.

• Development of several transaction controls to comply with SEPA rules.

• Creation of a SWIP (single window instant payment) transaction routing application, allowing STET to be the single entry point for transactions coming from multiple CSMs (Scala, Kafka Streams).

• Generation of statistical reports for Instant Payment transactions (Scala, Spark, Cassandra).

• Anomaly analysis and production of deliverables.

• Integration testing on a stack of VMs in AWS.

Technical Environment: Agile Scrum, Scala, Spark, Kafka, Kafka Stream, Cassandra, AWS, Jenkins, Git, GitHub, IntelliJ, Maven, Jira, Confluence, ScalaTest, Linux