Beelzebub is hiring for a team member at Agenzia per la Cybersicurezza Nazionale.

Attackers now move at machine-speed. They use frontier AI to find and weaponize vulnerabilities automatically, generate malware from natural-language prompts, and run agentic attacks faster than any human team can respond. The result is an unfair fight: AI on offense against human defenders. Beelzebub exists to even the odds. Beelzebub is an AI-native cybersecurity platform that fuses red team and blue team into a single autonomous, machine-speed purple-team loop. Continuous adversary emulation finds the exploitable paths before attackers do. Runtime deception detects intrusions with zero false positives and contains them in milliseconds. Autonomous threat intelligence turns every intercepted attack into actionable defense in minutes. One closed loop, running at the speed of the threat. Born from more than three years of open-source development, Beelzebub is trusted by a global community of 2,000+ GitHub stars and 450+ weekly installs across 45+ countries, and by engineers at companies including Microsoft, Google, Cisco and Red Hat. It is a member of the NVIDIA Inception Program and of Anthropic's Cyber Verification Program. This is not theory. In December 2025, Beelzebub's technology captured the operation of the TeamPCP threat group live, work later credited by the Cloud Security Alliance, the same group now behind one of the largest software supply chain campaigns to date. Today Beelzebub is deployed by corporate, critical and government organizations that need to defend at machine-speed, with native alignment to NIS2, DORA and the EU AI Act. The era of AI versus human defense is over. Welcome to AI versus AI.