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.
An enterprise development team is transitioning a generative AI prototype from experimentation to a large-scale, enterprise production environment on Google Cloud. The organization requires a standardized, automated method to reliably provision and manage cloud infrastructure components—such as Vertex AI endpoints, datasets, and runtime environments—consistently across multiple regions while preventing configuration drift.
Which tool and approach should the organization use to achieve reliable and scalable infrastructure provisioning for Vertex AI?
In Vertex AI, a feature view is a logical collection of machine learning features materialized from a BigQuery data source into an online store. It periodically synchronizes feature values from data warehouses to make them available for low-latency inference.
Terraform is an open-source infrastructure as code (IaC) software tool that enables teams to define, provision, and configure Google Cloud resources using a declarative, configuration-oriented syntax. By defining infrastructure in version-controlled configuration files, organizations can manage Vertex AI resources programmatically.
google_vertex_ai_endpoint), datasets (google_vertex_ai_dataset), and workbench runtimes consistently across development, staging, and production environments.Transitioning generative AI systems from experimentation to enterprise-grade production requires deterministic, repeatable, and audited infrastructure management. Terraform eliminates manual errors in the Google Cloud console and ensures production environments remain reliable, scalable, and secure.
Vertex AI Studio safety filtering is a responsible AI capability that evaluates prompts and model responses against predefined safety categories such as hate speech, harassment, sexual content, and dangerous content. It allows developers to define confidence thresholds and filter harmful model outputs.
Few-shot prompting is a prompt engineering technique where a small set of input-output examples is included within the context window of a foundation model prompt. These examples guide the model's tone, format, and response structure without updating the underlying model weights.
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