Design Rules For Resilient AI On S3, DynamoDB, And Iceberg

AI workloads do not fail where people expect. They fail where resilience was never designed in. At AWS re:Invent, I spoke with my friend Michael Fasulo, Senior Director for Portfolio, AI Resilience and ResOps, Commvault to understand where AI systems on AWS are actually breaking and what teams should do differently.

This conversation was practical and grounded in what customers are seeing today. We covered

- Where resilience breaks first in real world AI workloads running on AWS

- What ResOps means in simple terms and why it is becoming essential for AWS customers running AI at scale

- How architects should think about recovery when AI spreads across S3, DynamoDB, and Iceberg based data lakes

- The design rules that make recovery easier instead of more complex

- The most important AI driven resilience capability Commvault is building into its AWS portfolio over the next year

If you are building AI on AWS and assuming resilience will take care of itself, this conversation is a good wake up call.

#data #ai #awsreinvent #aws #agents #awspartners #awscompetencypartners #agenticai #theravitshow