Nour JAOUADI

Project 1: Restitution of banking data (credits, third parties, etc.) requested by the SG group for reporting and cross-functional applications, in compliance with new banking regulations imposed by the ECB (European Central Bank) - Development of new business rules: Understanding of business needs, development of solutions, functional tests on data, optimization of jobs, realization of MEPs - Monthly production monitoring: Monitoring of job launches, querying and verification of data produced, automation of data checks, management and resolution of production bugs Project 2: Migration of a project based on internal databases to DataLake. This enables extractions to be produced based on a more extensive and complete history from the DataLake. These extractions represent unpaid transactions within the scope of securitized receivables - Collaboration with the functional team to understand the requirements and ensure that the migration solution meets them - Development of the project from scratch on Scala/Spark to take advantage of DataLake - Creation of jobs on Control-M for monthly production - Comparison with the current project in production (on SQL) to validate the new solution and put the project into production

Acceleration of the data cleaning industrialization process to guarantee good data quality for Data Scientists in project modeling, in an international context - Building automated data cleaners - Industrializing pipelines for the various countries on Azure Data Factory: Optimize processing with PySpark, adding sanity checks on input and output data - Development of a python package containing data transformation functions to accelerate the development and industrialization of country-specific cleaners: CI/CD pipeline on Azure, unit tests, MkDocs documentation - Development of a python-based package for building executable SQL queries on Snowflake. This tool automates the team's requests, makes them parameterizable and facilitates development by making notebooks easier to read

Migration of the alert generation engine against terrorist financing (FT), already inproduction on SAS, to a Big Data environment (in collaboration with the financialsecurity department) - Monitoring, improving and maintaining FT engine production on SAS pendingmigration - Migration of the FT engine to a Hadoop DataLake environment: Design of the newsolution, PySpark development, job scheduling via Airflow, comparative acceptancetest - Expand project to include legal entities (associations, SMEs ...)

Developing a predictive model to identify complex insurance contracts in inheritance inlife insurance. The aim is to be able to identify these contracts that will be complex tosettle after the client's death, in order to take actions earlier - Analysis and updating of the Succession Pathway over the year to track changesand evolutions: Understanding of business needs, data processing and cleansing,assessment of customer satisfaction, visualization of variables - Modeling the NPS (Net Promoter Score) for Inheritance: The NPS is an indicator formeasuring the recommendation of a company by a customer: Data processing,testing of different models (Logistic Regression, k-NN, Random Forest), businessdecision based on the performance of metric results - Analysing of the complexity of files in succession: Cleaning of beneficiary clauses infiles, Feature Engineering on beneficiary clauses, testing Machine Leaning models,Tuning of hyper-parameters to improve the performance of the model, use of dataresampling methods (Undersampling, Oversampling) to improve results

- Modeling and prediction of vehicle sales data using Machine Learning algorithms - Study and analysis of vehicle ad consultation data - Automation of results on an RShiny application