VPC Security ML Quiz

5 questions Pass: 70% +25 pts (50 with Pro)

Quiz covering ML Security

What you'll be asked

  1. Which of the following is a primary security benefit of deploying an ML inference service within a Virtual Private Cloud (VPC) rather than on the public internet?
  2. You have an ML training job running in a private subnet that needs to download large datasets from an S3 bucket. Which mechanism provides the most secure way to access this data without traversing the public internet?
  3. An ML model serving API is deployed in a private subnet. You need to allow external applications to call this API without exposing the instance directly to the internet. Which architecture is best suited for this?
  4. When securing ML pipelines, what is the primary difference between a Network ACL (NACL) and a Security Group?
  5. You are auditing a VPC configuration for an ML deployment. You find that the ML inference instances are in a private subnet but have a route to a NAT Gateway that is currently allowing all outbound traffic (0.0.0.0/0). Why is this a potential security risk for ML model integrity?

Answers and explanations come after you submit.