ShieldAI – Intrusion Prevention System

Abstract

In today’s world of internet, safeguarding network infrastructure against malicious activities is crucial. This project is real-time threat detection that leverages machine learning and AI to detect and classify cyber-attacks based on live packet capturing. Utilizing Scapy for packet capturing and FastApi as a backend, the system extracts 71 features from packet and feeds them into a pre-trained Autoencoder and DNN based model to accurately predict potential threats such as DDoS, Port Scans, Brute Force, SQL Injection, and more. The detection results are streamed to a dynamic React- based frontend providing real-time threat visualization and monitoring. The system handles such attacks by IP blocking for malicious sources. So, this system will enhances proactive defense mechanisms and provide scalable solution for modern cybersecurity needs.