Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
A company is deploying proprietary machine learning models and sensitive prediction pipelines on Vertex AI in Google Cloud. The security architecture team needs an infrastructure safeguard that establishes a security boundary around the Vertex AI resources, prediction endpoints, and storage buckets to isolate model deployments and mitigate the risk of data exfiltration.
Which architectural security capability should the team configure to create this boundary?
The Service Account User role (roles/iam.serviceAccountUser) allows a user or automated pipeline to attach a service account to a compute resource, such as running a training job or deploying a model endpoint under the identity of that service account.
Using dedicated service accounts is an essential best practice for least privilege, but securing deployments against extraction and exfiltration requires encapsulating those accounts and workloads inside VPC Service Controls.
The Storage Object Viewer role (roles/storage.objectViewer) is an Identity and Access Management (IAM) role that provides read access to view and download object data and metadata within Google Cloud Storage buckets.
roles/storage.objectViewer gives read permissions to users or service agents, but it does not isolate Vertex AI prediction containers or secure inference API endpoints.storage.objects.get, storage.objects.list) to identities.IAM policies define the principle of least privilege for identities, but comprehensive ML security requires combining IAM with perimeter controls like VPC Service Controls to enforce defense-in-depth.
The Restrict Resource Service Usage constraint is an Organization Policy mechanism that allows cloud administrators to govern which Google Cloud APIs and services can be activated or consumed within specific resource hierarchy folders or projects across the enterprise.
aiplatform.googleapis.com) can be used in a project, not how the deployed endpoints are isolated at runtime.Organization policies provide macro-level governance by restricting service activation, whereas VPC Service Controls provides micro-level isolation and egress filtering for active services.
VPC Service Controls is a Google Cloud security technology designed to establish granular perimeter security around sensitive Google Cloud resources and managed multi-tenant services such as Vertex AI, Cloud Storage, and BigQuery. It enables organizations to define perimeter boundaries that isolate resources and restrict API communications strictly to authorized networks and identities.
While identity and access management controls verify who is making an API call, VPC Service Controls verifies where the request originates and where data can travel. This dual-layer strategy ensures that even if credentials are compromised, model artifacts, datasets, and endpoints remain protected inside the secure perimeter boundary.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated