My research focuses on applying modern Deep Learning in Cybersecurity, developing advanced techniques to support companies and public administrations in software security certification. Specifically, I work on implementing LLM-based vulnerability detectors and preventing adversarial attacks on LLMs.