Confidential AI with Ubuntu on Azure - A Deep Dive

Learn about confidential AI implementation on Azure through a technical webinar that explores how enterprises can protect sensitive data privacy and model intellectual property during cloud-based machine learning tasks. Discover how hardware-rooted execution environments spanning both CPU and GPU enhance protection of AI data and code at runtime, safeguarding against privileged system software and cloud operators. Explore the solution built on Ubuntu 22.04 confidential VMs using AMD 4th Gen EPYC processors with SEV-SNP and NVIDIA H100 GPUs. Delve into core components including memory confidentiality, integrity across CPU and GPU, and the role of remote attestation in securing AI workloads. Gain insights from Canonical's Ijlal Loutfi and Microsoft Research Cambridge's Antoine Delignat-Lavaud as they address how confidential AI helps organizations comply with industry regulations while maximizing the potential of private data in AI applications.