From Edge to Insight - Cybersecurity Intelligence with Kubeflow, VLLM, and GraphRAG

Explore a comprehensive conference talk that demonstrates building a security-centric, federated AI architecture using Kubeflow for modern cybersecurity challenges. Learn how to implement adaptive, policy-aware AI systems capable of autonomous threat detection, reasoning, and response across cloud and edge environments. Discover the deployment of lightweight ONNX models on K3s-based edge clusters for local Falco log analysis and anomaly detection while maintaining minimal exposure. Understand how to ensure compliance with GDPR, HIPAA, and zero trust principles by sharing only filtered signals and key features between edge and cloud components. Master the orchestration of GraphRAG pipelines in the cloud using KubeFlow, powered by vLLM and Knowledge Graph reasoning aligned with NIST and MITRE ATT&CK frameworks. Gain insights into implementing intelligent recommendation systems and rule automation capabilities, including policy and rule updates deployable back to edge environments. Acquire practical tools and methodologies specifically designed for high-stakes environments such as healthcare, operational technology, and critical infrastructure, enabling security and DevSecOps teams to deploy self-improving, compliance-aligned AI defense systems.