Derek Chamorro

Placeholder session for confirmed speakers without an assigned session, allowing them to be pushed to Swapcard and Visit.

Traditional SIEM platforms excel at detecting network intrusions and OS-level threats, but they're fundamentally blind to attacks living in GPUs, BMCs, and out-of-band management planes. As AI/ML workloads scale, attackers increasingly exploit this hardware-layer blind spot to steal models, achieve persistent access via firmware implants, and maintain presence even after incident response. This talk treats GPU and BMC telemetry as first-class forensic data, not just performance metrics, and turns them into concrete detections that a SOC can deploy. We'll cover GPU telemetry (DCGM), BMC logs (IPMI/Redfish), out-of-band monitoring and show how to operationalize them at scale across multi-vendor environments. You'll see the challenges of integration (protocol chaos, network isolation, and data volume) and a three-layer reference architecture (Collection ? Normalization ? Correlation), along with ready-to-implement rules and a realistic roadmap from week-one ingestion to behavioral baselining. Hardware-layer persistence is real. Your SIEM doesn't see it, yet. Learn what you're missing and how to fix it.