LLDP-IDS in SDN
Hybrid rule + RandomForest IDS that detects and mitigates LLDP topology-poisoning in software-defined networks at line rate.
Computer & Communications Engineering graduate from AUB, focused on blue-team security and DFIR, with work across machine learning, offensive security, cryptography, and semantic communications.
I'm a Computer & Communications Engineering graduate from the American University of Beirut, specializing in Cybersecurity & AI - High Distinction, Dean's Honors List every semester.
My focus is the blue team: Windows and Linux forensics, SOC monitoring and incident response, network forensics on real-world PCAPs, malware analysis, and reverse engineering. Beyond that, I work on offensive security, cryptography, and first-author research on semantic communication.
SECURITY · AIComputer & Communications Engineering graduate from AUB, specializing in Cybersecurity & AI. Focused on the blue team - Windows and Linux forensics, SOC and incident response, network forensics, and malware analysis - with offensive security, cryptography, and first-author research on semantic communication. IEEE GLOBECOM 2026, CTF finalist, Dean's Honors every semester.
Mada / CYMA Cybersecurity
American University of Beirut
American University of Beirut
B.E. Computer & Communications Engineering (Cybersecurity & AI)
Lebanese Baccalaureate - Mention Très Bien, Top of class
Hybrid rule + RandomForest IDS that detects and mitigates LLDP topology-poisoning in software-defined networks at line rate.
Blockchain-backed digital evidence platform: dual hot/cold Hyperledger Fabric, a private IPFS cluster, and mTLS/PKI across three hardened VMs. Final-year project.
Investigation of a multi-stage Windows intrusion: forensic timeline, lateral movement, persistence, full-packet network forensics, and MITRE ATT&CK mapping.
An on-path NFQUEUE DNS spoofer plus a passive detector that scores answers against authoritative resolvers and emits JSON alerts.
An Android cellular-survey app plus a Flask analytics backend: operator and network-type stats, geo heatmaps, and signal-quality prediction.
Compact deep semantic communication via Krum-inspired teacher selection and a residual channel bottleneck. IEEE GLOBECOM 2026, first author.
Deep joint source-channel coding with in-loop concept erasure (LEACE) inside a variational information bottleneck. IEEE GLOBECOM 2026, first author.
A slippage-aware arbitrage-cycle filter for DEX networks, with an adaptive-conductivity core derived from slime-mold (Physarum) dynamics. Solo journal submission.
Open a paper for details