AI models are increasingly critical to modern applications, yet most teams treat them as opaque binaries outside the bounds of traditional software supply chain security. This workshop guides participants through an end-to-end cloud-native pipeline that secures model artifacts from training to deployment. Using open tools like KitOps, Sigstore/Cosign, and Kubernetes, we’ll package a Hugging Face model, generate and verify attestations, enforce policies, and trace provenance. This is not just theory — it’s hands-on, practical, and designed to align with cloud-native workflows. Attendees will leave with patterns and tools they can immediately apply to secure AI in production.