
AI-powered intrusion and fraud detection system designed to enhance security within the Qubic Network. The system monitors real-time transaction data, analyzes wallet behavior, and detects suspicious patterns such as unusually large transfers, repeated failed transactions, and anomalous contract interactions. Using a combination of rule-based detection and machine learning (Isolation Forest), it assigns a risk score to each wallet and generates instant alerts via Discord or Telegram. Built entirely on Google Colab with Python and Streamlit, the prototype demonstrates how lightweight, autonomous security tools can protect DeFi ecosystems and improve user trust.
7 Dec 2025

AI-driven Intrusion Detection System designed to detect suspicious network activity in real time. Modern networks face continuous cyber threats such as brute-force login attempts, port scans, unauthorized access, and malware activity. Traditional IDS tools struggle with new attack patterns and generate many false alerts. Our solution combines Machine Learning classification, a rule-based detection engine, and an LLM explanation layer to provide accurate, understandable threat detection. Users can submit network logs through a real-time Flask API hosted via ngrok. The system classifies events, checks for known attack signatures, and generates a human-friendly explanation of the threat. This solution targets SMEs, IT teams, and developers needing lightweight, affordable, and API-friendly network protection.
19 Nov 2025