GCP Professional Cloud Architect Practice Question
Your company wants to launch a chat-based help-desk assistant within two weeks. The backend runs on Cloud Run and must call the assistant through a private REST endpoint. The team has no capacity to train or maintain ML serving infrastructure but wants the option to later refine the model with proprietary support-chat logs without changing application code. Which approach best meets these requirements with minimal operational overhead?
Download the model artifacts from Model Garden to Cloud Storage, build a custom serving container, and deploy it on a GKE Autopilot cluster behind an Internal HTTP(S) Load Balancer; fine-tune by retraining the model and rebuilding the container.
Package the model code inside a Cloud Function that serves HTTP requests; when fine-tuning is required, upload new model weights to Cloud Storage and redeploy the function with updated code.
Use Vertex AI Model Garden to select a suitable foundation model, click "Deploy to Vertex AI" to create a managed prediction endpoint, invoke it from Cloud Run via Private Service Connect, and later apply Vertex AI tuning to generate a new model version that can be redeployed on the same endpoint.
Deploy an open-source LLM from Cloud Marketplace on a managed instance group in Compute Engine and expose it through an internal TCP/UDP Load Balancer; perform future fine-tuning by running custom training jobs on separate GPU VMs.
Vertex AI Model Garden lets you browse Google-hosted and third-party foundation models, test them in the console, and deploy any supported model directly to a managed Vertex AI endpoint with a single click or API call. The endpoint is fully managed-scaling, patching, and serving infrastructure are handled by Google, so the team does not need to run GKE or custom VM images. When they are ready to improve quality, they can invoke Vertex AI's built-in tuning service (e.g., adapter-based or full fine-tuning) to create a new tuned Model resource and redeploy it to the same endpoint without changing the calling code used by Cloud Run. The other options all require the team to build, deploy, and operate custom serving infrastructure or to redeploy application logic when the model changes, which violates the stated constraints.
Ask Bash
Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.
What is Vertex AI Model Garden?
Open an interactive chat with Bash
What is Private Service Connect and how does it enable secure communication?
Open an interactive chat with Bash
How does Vertex AI's model tuning work?
Open an interactive chat with Bash
What is Vertex AI Model Garden?
Open an interactive chat with Bash
How does Private Service Connect enable secure communication in this scenario?
Open an interactive chat with Bash
What is adapter-based fine-tuning in Vertex AI?
Open an interactive chat with Bash
GCP Professional Cloud Architect
Managing and provisioning a solution infrastructure
Your Score:
Report Issue
Bash, the Crucial Exams Chat Bot
AI Bot
Loading...
Loading...
Loading...
Pass with Confidence.
IT & Cybersecurity Package
You have hit the limits of our free tier, become a Premium Member today for unlimited access.
Military, Healthcare worker, Gov. employee or Teacher? See if you qualify for a Community Discount.
Monthly
$19.99 $11.99
$11.99/mo
Billed monthly, Cancel any time.
$19.99 after promotion ends
3 Month Pass
$44.99 $26.99
$8.99/mo
One time purchase of $26.99, Does not auto-renew.
$44.99 after promotion ends
Save $18!
MOST POPULAR
Annual Pass
$119.99 $71.99
$5.99/mo
One time purchase of $71.99, Does not auto-renew.
$119.99 after promotion ends
Save $48!
BEST DEAL
Lifetime Pass
$189.99 $113.99
One time purchase, Good for life.
Save $76!
What You Get
All IT & Cybersecurity Package plans include the following perks and exams .