Secure your healthcare generative AI workloads on Amazon EKS
summary
Healthcare generative AI on Amazon EKS is presented as a deployable pattern that combines BioMistral‑7B from Hugging Face with Ray/KubeRay for scaling, a FastAPI or Gradio front end, and Carpenter node scaler for just‑in‑time accelerated nodes; threat modeling uses STRIDE and cites a surprising 94% of executives insisting on pre‑deployment security. Recommended controls include RBAC + IAM least‑privilege, CloudTrail audit logs, signed images (AWS Signer/notation), image scanning, KMS/Secrets Manager envelope encryption, network policies, GuardDuty runtime monitoring, and resource limits to shrink blast radius.
AI-written summary · about
chapters (11)
- 0:00Introduction and goals
- 1:16GenAI in healthcare
- 4:23The Stride threat model
- 5:41Security threats detailed
- 8:31Intelligent health assistant
- 8:54Architecture and deployment
- 9:26Mitigation strategies
- 17:46Security best practices
- 17:59Reference blueprints
- 18:09Closing remarks
- 19:04Conclusion
related talks
27mGen-AI at Scale: Simplifying Orchestration of Healthcare Applications Across Multi-Cluster Environme
26mGenerative AI driven sensitive data protection
35mBuilding a Medical AI with Kubernetes and Kubeflow
25mSecuring AI Pipelines: Real-World Attacks on Kubernetes-Based AI Infrastructure
45mGenerative AI in medicine: Improving patient & provider experiences
40mSecure Your GenAI Applications
35mHybrid-Cloud, HIPAA Compliant Enterprise with Kubernetes
09mSecuring Generative AI Workloads