Skander Malouche

Within ateam, composed of five data scientists. The team’s main responsibilities include developing pricing optimization models for consumer credit that maximize bank revenue, as well as conducting marketing-driven studies focused on customer insights and profiling: Refactoring and Enhancement of Internal ML Code: Refactored, adapted, and improved a Python-based Machine Learning modeling package (internally developed libraries), aligning it with current market needs and the bank’s commercial strategy. The project focused on enhancing the internal ML package, designed to provide personalized loan rates to clients applying for consumer credit—offering interest rates that balance customer suitability with optimal profitability for the bank. Business Rebate Data Analysis: Analyzed rebate (reimbursement) data for business clients, enabling the commercial team to design targeted repayment strategies that reduce bank costs. Tasks involved constructing KPI variables using customer knowledge databases (SAS) and performing analysis through Excel pivot tables. Tools: Python, Domino, SQL (via SAS), Excel

Within the Data & Artificial Intelligence department, the Retail team of 25 data experts focuses on data science challenges related to marketing for both individual and business market segments: Real-Time AI & Dynamic Personalization: Developed a Machine Learning model to predict, in real time, a customer’s propensity to subscribe to a banking product while browsing the SG website or mobile app via digital channels. Intelligent Segmentation & Profiling: Created Machine Learning models for customer profiling and segmentation to build subscription personas for banking products. Advanced Predictive Modeling with Supervised Learning: Developed an attrition scoring model for retail customers to anticipate churn using algorithms such as XGBoost and Random Forest. AI-Assisted Product Recommendation Engine: Deployed a banking product recommendation engine used by advisors to suggest products to clients, leveraging a hybrid approach combining collaborative filtering and supervised learning models. Detection of Competitive Commercial Opportunities: Implemented a detection model for identifying specialized financing obtained by clients through competitor institutions. Customer Lifetime Value Estimation: Built a Customer Lifetime Value model to estimate potential profit a customer could bring to the bank, including future value projection based on financial data. Feature Store Development: Designed and implemented a centralized feature store to standardize and share features across data science projects, enhancing reusability and efficiency. R&D in Generative AI: Led R&D initiatives in Generative AI (Large Language Models, fine-tuning, prompt engineering) to develop AI assistants tailored to data science needs Tools: PySpark, Spark, H2O, Python, SQL, Hadoop, Hue, GIT, Dataiku

Mission: Video analysis, person recognition and labeling by assigning tags for the video sequences via deep learning models managed by Amazon Web Services cloud services. Tools: AWS (SageMaker, Amazon Rekognition, S3, IAM, SNS, SQS), Python.

Missions: Creation of a product recommendation system developed on Big Query GCP based on the history of customer transactions using support, lift and confidence metrics. Centralization of data from Big Data Hadoop platform and GCP via the creation of a global product repository. Refactoring scripts adapting them to new data sources. Tools: PySpark, SQL on Big Query, Google Cloud Platform GCP, Google storage Git, Bitbucket

Mission: Refactoring of an automatic product pricing project from a version coded in R to Python and adapting it to the new Azur databases and to current business needs. Tools: Python, R, SQL, Git, Jira. Work environment in agile mode and in English.

Mission: Development of customer retention’s scoring models for an automobile insurance product using different machine learning algorithms as Generalized Linear Models, Penalized Regressions, Random Forest, XGBoost. Tools: R, SAS.

Department: Statistics department (StatSC : Staistique Sensométrie Chimiométrie) of Oniris engineering school, Mission: Segmentation and analysis of consumer preferences for food products using clustering methods based on neural networks. Tools: R, Excel.

Department: Observatory of the economic situation, Main mission: Processing of seasonal time series involved in the calculation of GDP, modeling, forecasting, correction of seasonal variations. detecting calendar effect and stationarity testing. Tools: Eviews, Jdemetra, R

Entreprise: International Arab Bank of Tunisia (BIAT), Department: Operational risks management department, Main mission: Development of a risk management model using an advanced measure approach: LDA. Tools: R, Excel

Department: Direction of commercial network. Main mission: Detection of anomalies in related entries from the client folder.

Department : Observatory of the economic situation. Main mission: Realisation of a descriptive study and hypothesis tests on a time series of the Industrial Production Index. Tools: R, Excel.