Securing Private LLMs and Data Center Infrastructure Without Sacrificing GPU Performance
Anchor: Register
To maintain data sovereignty and protect core intellectual property, organizations are rapidly moving AI workloads into private environments. Unfortunately, legacy security systems fail when facing massive data flows and novel AI threat vectors. In this compact, 30-minute briefing, Check Point architecture experts deliver a comprehensive framework to harden the modern AI factory. You will discover how to build a multi-layered defense across the entire architecture, securing everything from perimeter prompt inspection and real-time inference traffic down to the physical GPU silicon and application layer. ─────────────────────────────────── This 30-minute webinar covers:
The AI Data Center Architectural Shift: Why the modern AI factory demands a new security paradigm to accommodate assets that are autonomous, semantically manipulable, and capable of cascading action.
Inference Edge & API Hardening: Intercepting adversarial queries, prompt injection, and manipulation at the ingress traffic layer.
Offloaded Hardware-Accelerated Security: Running inline threat prevention natively on NVIDIA BlueField DPUs without any impact on AI GPU performance
Kubernetes Cluster Isolation: Enforcing east-west micro-segmentation at the runtime level to block lateral threat movement between tenant inference namespaces.
Agentic Layer Access Control: Applying dynamic Zero Trust validation and data exfiltration filtering to autonomous AI system calls and LLM endpoints.
Subtitle 1
Subtitle 2
Subtitle 3
Speakers
Avi Rembaum
President, Technical Sales Check Point Software Technologies
Jeff Schwartz
Vice President, Americas Engineering Check Point Software Technologies
