Your organization is building a real-time recommendation engine using ML models that process live user activity data stored in BigQuery and Cloud Storage. Each new model developed is saved to…

Question

Options

• AEnable container image vulnerability scanning during development and pre-deployment. Enforce
• BThoroughly sanitize all training data prior to model development to reduce risk of poisoning
• CLimit external libraries and dependencies that are used for the ML models as much as possible.
• DDevelop strict firewall rules to limit external traffic to Cloud Run instances. Integrate intrusion

• A
  79% (45)
• B
  12% (7)
• C
  7% (4)
• D
  2% (1)

Explanation

ML model supply chain attacks target the pipeline used to build, store, and deploy models. Since models are saved as container images in Artifact Registry and deployed to GKE, the attack surface mirrors a standard container supply chain. Enabling vulnerability scanning in Artifact Registry during development and pre-deployment catches compromised or vulnerable dependencies, and enforcing Binary Authorization ensures only cryptographically attested (verified) images can be deployed to GKE - directly closing the supply chain attack vector. Option B (sanitizing training data) addresses data poisoning attacks, not supply chain compromise of the deployment pipeline. Option C (limiting libraries) reduces attack surface but provides no automated enforcement or verification. Option D references Cloud Run firewall rules and intrusion detection, but the architecture uses GKE, not Cloud Run, and these controls address network threats rather than supply chain risks.

No community discussion yet for this question.