Douaa EL-YOUSRI

– Scraping and integrating over 22,000 property listings from Avito and Mubawab (BeautifulSoup, Selenium) – Cleaning, harmonizing, and merging data to build a unified and reliable dataset – Performing Exploratory Data Analysis (EDA) and district segmentation using K-Means (elbow method and silhouette score) – Building predictive models with Linear Regression, Lasso, Ridge, Random Forest, Gradient Boosting, LightGBM, and XGBoost (R² ≈ 0.92, RMSE ≈ 256K MAD) – Deploying a FastAPI endpoint serving the final model, connected to a Next.js web interface for automated price estimation – Visualizing and monitoring results and KPIs in Power BI

The study employs an advanced combination of Machine Learning models, including K-means, Isolation Forest, SVM One-Class, and Local Outlier Factor (LOF), to detect fraud in public procurement. These techniques were utilized to analyze anomalies in a complex dataset comprising information on contract amounts, the number of bidders, and market types. Additionally, a consensus method was implemented to merge the results of the different models and enhance detection reliability. Performance evaluation, conducted using metrics such as precision and anomaly detection rates, demonstrated the effectiveness of these approaches in addressing the specific needs of the General Treasury of the Kingdom. This project thus contributes to strengthening control mechanisms and improving transparency in public finance management.

ENSIAS is one of Morocco’s top engineering schools, recognized for preparing its students to excel in various fields of computer science and digital technologies. The school provides high-quality education that enables students to tackle the challenges of a rapidly evolving technological world. Core Curriculum: - Algorithms and Data Structures - Programming in C, Linux, and Object-Oriented Programming (OOP) - Applied Probability, Simulation, and Statistics - Databases (SQL & NoSQL) - Analysis and Operations Research Specialization: - Data Analysis and Data Science - Machine Learning and Neural Networks - Stochastic Modeling in Finance - Investment Techniques and Ethics in Finance - International Finance - Advanced Optimization Techniques

An excellent training within one of the top preparatory classes in France. This experience allowed for the development of in-depth skills in various areas: Analysis, algebra, and probability: Advanced study of fundamental mathematical concepts essential for modeling and solving complex problems. Computer science: In-depth exploration of algorithms and data structures, with proficiency in Python and OCaml. Intensive preparation for the entrance exams to the Grandes Écoles, combining academic rigor, teamwork, and pressure management.