Education

Politecnico di Torino logo

Master of Engineering, Cybersecurity Engineering

Politecnico di Torino - Turin, Italy

Sep 2024 - Oct 2026

GPA 29.48/30

MSc Cyber Security Engineering (LM 32). English-taught, ECSO-aligned curriculum.

My training is grounded in applied security research, not only theoretical coursework.

  • Kernel vulnerability research: I performed a deep security analysis of Linux kernel 6.8 eBPF verifier logic and developed PoC exploits for 5 confirmed vulnerabilities.
  • AI for cybersecurity: I engineered a multi-model threat detection suite and reached 98.1% malware classification accuracy using GNN, RNN, and FFNN models.
  • NLP-based detection: I built a MITRE ATT&CK tactic classifier with BERT and UniXcoder to map bash command sequences to adversary tactics.
  • Post-quantum cryptography: I am researching post-quantum secure policy-based anonymous authentication tokens for my thesis at Huawei (FIPS 204/205).
Politecnico di Torino city and campus context
Universita degli Studi del Sannio logo

Bachelor of Engineering, Computer Engineering

Universita degli Studi del Sannio-Benevento - Benevento, Italy

Sep 2021 - Jun 2024

101/110

BSc in Computer Science Engineering (L-8).

Thesis research: Extracting Equivocal Behaviours from Trusted Systems.

  • Sandbox orchestration engineering: I designed and implemented an asynchronous multithreaded Python orchestrator to coordinate Hybrid Analysis, VirusTotal, and ANY.RUN sandbox analysis across Windows 10 and 11.
  • Behavioral taxonomy mapping: I defined and mapped 12 Equivocal Software Behaviours directly to the MITRE ATT&CK Enterprise Matrix.
  • Detection outcome: I detected system reconnaissance and advanced OS utility exploitation in 100% of analyzed trusted systems, exposing silent telemetry and hidden attack vectors.
Unisannio city and course context
Zenodo publication record

Extracting Equivocal Behaviours from Trusted Systems

My capstone addressed software supply chain attacks by identifying undocumented or malicious behaviours hidden in trusted software and validating findings with reproducible sandbox-based analysis.