Saoud's Projects

A machine learning web app that predicts whether a Titanic passenger would have survived. Enter passenger details like class, age, and gender to get an instant prediction with survival probability.

A role-based document management system with four clearance levels — Unclassified, Classified, Secret, and Top Secret. Features file upload, dark mode, Arabic/English support, and strict access control per user role.

A professional cybersecurity services webpage offering tailored security solutions, penetration testing, threat analysis, and consulting services for individuals and businesses.

An advanced phishing detection tool that analyzes URLs, IP addresses, domains, and file hashes using heuristics, Google Safe Browsing, and 70+ VirusTotal antivirus engines.

An autonomous AI agent that acts as a Tier-1 SOC analyst. Uses a coordinator-mainagent + 3 specialist subagents with 10 investigation tools to classify security alerts as Malicious, Suspicious, or Benign. Purely agentic — no hardcoded rules. 100% accuracy on 20 alerts, 95.7% on 600 alerts. Supports cloud (OpenRouter) and local (Ollama) LLMs.

A machine-learning-based Network Intrusion Detection System trained on the CIC-IDS2017 dataset. Classifies network flows as Benign or attack types (DDoS, DoS, PortScan, Bot, BruteForce, WebAttack). Built with XGBoost, PyTorch, and Streamlit.

Browse all my open-source projects and source code on GitHub, including SOC Deep Agents, NIDS, PhishRadar, Titanic Survival Predictor, and more.