Training jobs and the production endpoint still start with default-VPC public IPs. Staff need to isolate the ML system, not attach an already-designed VPC. What should they design?
Select an answer to reveal the explanation.
Short Explanation
Jobs and the endpoint still start with default-VPC public IPs. Design private subnets, routing, and security-group boundaries. Default-VPC public IPs and an internet-wide rule are the problem, not the design.
Full Explanation
Isolating ML systems means designing private subnets, routing, and security groups so jobs and endpoints are not on the public internet. Default-VPC public IPs and an internet-wide rule are the problem, not the design. Rekognition is not a subnet plan.