George Kurtz

Kurtz’s owned biography describes joining Price Waterhouse’s security division in 1993. Foundstone combined vulnerability-assessment products with incident response and training before McAfee acquired it. At McAfee, his responsibilities expanded from general management to technology leadership. He later worked with Warburg Pincus as an entrepreneur in residence while developing CrowdStrike’s cloud-based approach.[3]

In the joint SafeMind announcement, CrowdStrike and NVIDIA describe a system pairing security models with an agent harness. A simulated environment lets offensive and defensive agents challenge each other: the defensive side observes sensor information, proposes detections, validates candidates and promotes them. NVIDIA’s account distinguishes an orchestrating model from a rule-generation model and describes post-training on CrowdStrike’s threat data. The contribution is a company system announced by Kurtz and Jensen Huang, with task controls and feedback around the models.[6]

In the TBPN conversation, Kurtz and the other participants discuss matching specialized open models and frontier models to different security tasks. They connect inference cost with the need to inspect a production environment continuously, including configuration as well as code. The conversation makes the operating problem concrete: the system needs a useful response quickly enough, at a cost compatible with repeated monitoring.[5]