Anis T.

Within the framework of the ACPR regulatory audit of the Generali Group, a large-scale data transformation program was launched to strengthen regulatory reporting, improve data quality, and implement a modern data mesh architecture.

Main contributions:

➔ Designed and implemented data ingestion, cleaning, and normalization pipelines within the central Data Lake (Dataiku, PySpark, Python, SQL).

➔ Built and automated datamarts and performance indicators to meet ACPR regulatory requirements.

➔ Structured workflows by business domains (“Flow Zones”) and managed inter-process dependencies to ensure optimal coordination.

➔ Designed KPI dashboards and monitoring tools for operational and regulatory reporting (Power BI).

➔ Developed data preparation, enrichment, and segmentation processes for Data Products (Snowpark, Python, SQL).

➔ Implemented data quality controls, automated jobs, and monitoring systems (Airflow, Splunk, Power BI).

➔ Ensured secure access management and secret handling (Snowflake, Rancher, Kubernetes).

➔ Set up CI/CD pipelines and DAG orchestration for automated deployments (Airflow, Azure AKS, ACR).

➔ Defined and applied GitFlow development standards and contributed to best practice sharing.

➔ Authored technical and functional documentation (Jira, Confluence) and supported strategic data-driven decision-making.

Tools:

Azure (AVD, AKS, ACR, ADLS, ADF), Snowflake, Snowpark, Dataiku, Python, Jupyter, Anaconda, Hadoop, PySpark, Airflow, Kubernetes, Rancher, Docker, GitLab, Splunk, Power BI, Jira, Confluence.

The aim of the project was to establish an automated platform for the design, production, and administration of Data Science models, along with the management of Data Catalog, Data Privacy, Data Quality, and Reporting.

Responsibilities included:

➔ Designed and implemented a platform for the creation and industrialization of Data Science models using Sagemaker.

➔ Implemented DevOps architecture.

➔ Organized datalab S3 buckets and managed exchanges with different AWS services.

➔ Assisted the data science team in using the AWS environment.

➔ Conducted a proof of concept and implemented pipelines for the ingestion, historical storage, and processing of Power-BI activity logs and SAS extractions within the datahub.

➔ Implemented databases in the Consumer Zone and connected them to PowerBi (Infra as Code, network flow openings, AWS routes, ODBC sources, data connections).

➔ Built the datacatalog of the datalab account and developed a POC allowing access to this data via Power-BI (Athena, Crawler, RDS, S3).

➔ Established workflows and processes for non-event-based data ingestion (Step Function, Lambda, Python).

Tools:

AWS (MSK, EMR, Glue, Athena, Redshift, Lambda, DynamoDB, RDS, S3, Step Function, CloudWatch, EC2, SSM, SNS, SQS, etc ) Scala, Spark, Kafka, Grafana, Python, Conda, Jupyter, ML, AirFlow, Docker, Kubernetes, Scrum, Jira.

Within the SGIB cybersecurity service, the objective was to implement a platform of intelligent controls to detect fraud. The main controls include detecting data leaks from the bank's information system through browsing log analysis, identifying fraudulent emails, detecting suspicious users attempting to leak data from the system, identifying risk profiles through the fusion of results obtained by various internal controls, verifying compliance of shared information system file owners, and analyzing authentication logs to identify suspicious or fraudulent connections.

Responsibilities included:

➔ Designed and implemented data ingestion and processing jobs using Big Data technologies.

➔ Developed and optimized Data Science processes for performance through parallelization and scaling.

➔ Trained Machine Learning models and established automatic retraining.

➔ Industrialized and automated fraud detection controls as infra-as-code in a DevOps environment.

➔ Updated controls to accommodate platform updates and upgrades for Big Data tools.

➔ Established a cross-platform monitoring solution for proactive incident resolution.

➔ Developed an API for managing referential data linked to cybersecurity.

➔ Migrated jobs and controls to the new BigData platform on Azure, providing daily support for production and execution monitoring.

Tools:

Microsoft Azure (ADLS, HDinsight, VDI, AKS, Azure Cli, Azure Container Registry, etc), HDP (HDFS, Yarn, Spark, Hive, Oozie, Ranger, Livy, Zeppelin, Kerberos), Anaconda, Python, ML, AirFlow, Docker, Kubernetes, Spring-Boot, Java.