Mohamed Elmouden

Intrusion Detection System for Software-Defined Networks using Machine Learning. Achieved 99.64% accuracy on IoTID20 dataset and 97.41% on CICIDS2017 using models like Random Forest, Decision Tree, and deep learning approaches.

Technical documentation assistant built for Orange Business using Retrieval-Augmented Generation. Leverages Azure OpenAI and LLMs to provide instant, accurate answers from vast document repositories via FastAPI microservices.

Distributed system architecture using Spring Boot and Spring Cloud. Features separate services for products, recommendations, and reviews with a central API Gateway, Eureka service discovery, and Docker containerization.

Deep learning project for medical imaging diagnosis. Trained VGG16, InceptionResNet, and custom CNN models on medical imaging data with image augmentation. Built an interactive Streamlit interface for predictions.

Custom simulation environment based on OpenAI Gym. Implementation and performance comparison of reinforcement learning algorithms (A2C, PPO, DQN) across different configurations.

Full-stack application for medical clinic client management. Implemented CRUD operations for patients and consultations with Spring Boot backend, Spring Security authentication, and a modern frontend.

Big data classification project using PySpark on Azure Databricks. Trained 5 models (Decision Tree, Logistic Regression, Random Forest, OneVsRest, Naive Bayes) with performance comparison across cluster configurations.

Creative AI application that automatically generates children's stories in text, image, and audio formats using artificial intelligence. Integrates multiple AI modules with a Spring Boot backbone.