最新のProfessional-Agentic-Architect問題集で試験を準備する

You are designing an enterprise agentic system in Google Cloud. The workload includes two distinct microservices: a stateful orchestrator that manages multistep reasoning workflows with asynchronous tool execution loops and agent-level security boundaries, and a stateless transformation service that handles unpredictable, bursty data processing workloads. You need to deploy these services using managed Google Cloud runtimes while minimizing operational management overhead and infrastructure cost.
What should you do?

• A. Deploy the orchestrator service on Agent Runtime, and host the transformation service in a Code Execution sandbox.
• B. Deploy the orchestrator and the transformation service on Agent Runtime as unified agent capability endpoints.
• C. Deploy the transformation service on Agent Runtime, and host the orchestrator on Google Kubernetes Engine (GKE) Autopilot clusters.
• D. Deploy the orchestrator on Agent Runtime for managed workflow lifecycles, and deploy the transformation service on Cloud Run.

You manage a support agent that uses RAG to answer questions based on a document corpus in RAG Engine.
Each query retrieves the top 50 chunks. While reviewing a set of failing questions, you determine that the chunk containing the answer is among the 50 returned but is ranked near the bottom, and the agent answers from a higher-ranked chunk instead.
You want the answering chunk to rank at the top without re-indexing the corpus. The retrieval step has 200 milliseconds of headroom.
What should you do?

• A. Add the Agent Platform ranking API as the Agent Search reranker to reorder the chunks.
• B. Add a reranker that calls a Gemini model to score each retrieved chunk, and reorder the chunks based on the user's input.
• C. Update the RAG Engine deployment configuration from the Spanner mode Basic tier to the Scaled tier.
• D. Set a vector_distance_threshold on the RAG Engine retrieval configuration to return more relevant chunks.

You are building an automated medical research assistant agent in Gemini Enterprise to analyze patient records for rare diseases. Ingestion includes high-resolution radiology scans and unstructured handwritten clinical notes stored in Cloud Storage. The agent must correlate fine visual tissue abnormalities in the radiology scans with contextual annotations in physician notes.
You need to configure the data ingestion and modeling pipeline to maximize diagnostic correlation without losing spatial or visual details.
What should you do?

• A. Extract image feature embeddings using a separate vision-only classification model, and store them in Vector Search along with plain text notes.
• B. Send the raw Cloud Storage scan and note URIs to the native multimodal Gemini model using full- resolution file references.
• C. Process the notes using an external optical character recognition (OCR) service, and downsample radiology scans to low-resolution JPEG images.
• D. Convert the radiology scans into text-based pixel matrix arrays, convert the notes to text, and append both into standard text prompt templates.

You are designing a continuous evaluation pipeline for an autonomous agent deployed on Agent Runtime.
The agent uses multiple valid, non-linear sequences of function calls to resolve user requests rather than hardcoded golden trajectories. You want to continuously assess the semantic validity of tool selection and parameter generation without using hardcoded reference files.
What should you do?

• A. Configure the pipeline to use the managed types.RubricMetric.FINAL_RESPONSE_QUALITY metric in the gen AI evaluation service.
• B. Export the agent's intermediate tool calls to BigQuery, generate vector embeddings using the ML.GENERATE_EMBEDDING function, and execute the VECTOR_SEARCH command.
• C. Configure the pipeline to use custom function metrics in the gen AI evaluation service.
• D. Configure the pipeline to use the managed types.RubricMetric.TOOL_USE_QUALITY metric in the gen AI evaluation service.

You are setting up a custom agent in Gemini Enterprise that uses a dedicated service account to query an Agent Search data store. During testing, grounded search queries fail with a 403 Permission Denied error indicating that the principal lacks the discoveryengine.servingConfigs.search permission on the data store resource. You need to fix this permission error while following the principle of least privilege.
What should you do?

• A. Grant the roles/discoveryengine.editor IAM role to the agent's service account on the Google Cloud project hosting the data store.
• B. Grant the roles/discoveryengine.viewer IAM role to the agent's service account on the data store resource.
• C. Grant the roles/storage.objectViewer IAM role to the agent's service account on a Cloud Storage bucket containing the original source documents.
• D. Grant the roles/discoveryengine.agentspaceAdmin IAM role to the agent's service account on the target Gemini Enterprise app.