Zeyad Hassan Abdelghany

For my graduation project, I developed a hybrid machine learning-based Web Application Firewall (WAF) to detect SQL injection (SQLi) and Cross-Site Scripting (XSS) attacks. The project combined two models: BERT, for processing web request text, and SVM, for classifying attacks. I simulated web attacks using a website built with Node.js, React, and MySQL, collecting payload data labeled as "normal" or "anomalous" to train the models. To process the payloads, I used TF-IDF for feature extraction. The machine learning models were integrated into a Flask web application, allowing real-time detection of malicious traffic. The WAF successfully identified and blocked various attack attempts, improving web application security. This project demonstrated the effectiveness of combining machine learning models for threat detection and strengthened my skills in machine learning, web development, and cybersecurity.