AI-500 무료 덤프문제 온라인 액세스

You have a Microsoft Foundry multi-agent solution. The solution includes an orchestrator that uses Azure Al Search as a semantic cache to shortlist agents and invokes the top match.
You have a version-controlled repository that contains a curated dataset, expected routing, and expected responses. A pull request adds an agent named Agent 1, new sample utterances, and a changed selector prompt.
You need to configure an automated prerelease evaluation gate for the pull request. The solution must meet the following requirements:
* Validate the integrated system response after the new agent is available to orchestration.
* Validate the semantic cache retrieval quality against the expected results.
* Keep regression tests aligned with the pull request changes.
How should you configure the evaluation gate? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.

정답:

Explanation:
Semantic cache metric: Candidate-set recall and precision; Integrated system check: An end-to-end evaluation against the curated dataset; Regression dataset: Add cases for the new capability to the curated dataset.
The semantic cache is acting as a retrieval/shortlisting component, so recall and precision are the appropriate measures of whether expected agent candidates are surfaced and whether irrelevant candidates are avoided.
The pull request also changes the selector prompt and adds an agent, so component-only tests are insufficient; an end-to-end evaluation must exercise retrieval, routing, the new agent, and final response generation together. Finally, the version-controlled regression dataset should evolve with the capability. Adding representative cases for the new agent keeps expected routing and responses aligned with the code and prompt changes in the pull request. Microsoft Foundry evaluation guidance treats curated datasets as reusable regression assets for CI/CD quality gates. Together, these three selections test retrieval quality, integrated behavior, and regression coverage at the correct layers. The evaluation should also preserve correlation identifiers and version information where possible so a failed score can be traced back to the exact agent, model, tool call, or retrieval step that produced it. This turns the metric into an actionable diagnostic rather than only a dashboard number.
Official Microsoft reference: Microsoft Foundry - evaluation datasets and CI/CD evaluation

You have a Microsoft Foundry multi-agent solution that includes a planning agent and an execution agent You are designing a fixed order process where executions can run for several hours and can be interrupted by worker restarts.
You need to implement orchestration for the process. The solution must meet the following requirements
* Pause at predefined points.
* Resume from the last saved state after an interruption.
Which mechanisms should you use? To answer, select the appropriate options in the answer area.
NOTE: Each correct selection is worth one point.

정답:

Explanation:
Process structure: A workflow with agent executors; Pause mechanism: A human-in-the-loop (HITL) gate; Interruption recovery: Checkpoints.
The process has a fixed execution order and can run for several hours, so an explicit workflow with agent executors is preferable to free-form agent delegation. The requirement to pause at predefined points is a human-in-the-loop requirement: the workflow must be able to suspend and emit a request for approval or operator input. The requirement to resume after worker restarts is addressed by checkpoints. Microsoft Agent Framework checkpoints persist workflow progress, executor state, pending messages, shared state, and outstanding requests so execution can be restored from the last saved point. A retry policy alone re-executes operations but does not reconstruct the complete workflow state after a process interruption. The three selected mechanisms therefore cover structure, controlled pause, and durable recovery as separate but complementary workflow concerns. In production, add telemetry and regression tests around this behavior so changes to prompts, models, tools, or orchestration do not silently alter the intended contract. The selected approach is the one that best matches the platform ' s native execution semantics.
Official Microsoft reference: Microsoft Agent Framework - Checkpoints and Human-in-the-loop

You have a Microsoft Agent Framework workflow processor that receives customer support requests from a queue. Each request is evaluated by three independent Microsoft Foundry agents.
You discover that the current processor dequeues 30 requests at a time and starts all agent runs immediately.
During peak load, as many as 90 agent runs execute simultaneously, and the downstream API receives partial participant messages.
A single agent run completes in five seconds at the 95th percentile (95p), and the target throughput is 120 requests per minute.
You need to change the orchestration to ensure that it meets the throughput target and prevents more than 30 agent runs from executing simultaneously. The solution must produce one consolidated downstream payload for each request.
What should you do?

• A.Dequeue 10 requests per batch. Start 10 ConcurrentBuilder workflows that include the three agents as participants.
• B.Dequeue 30 requests per batch. Start a single ConcurrentBuilder workflow that includes 90 agent executors as participants.
• C.Dequeue 10 requests per batch. Start 10 group chat workflows that include the three agents and a manager.
• D.Dequeue 10 requests per batch. Start 10 sequential workflows that include the three agents as ordered steps.

You have a multi-agent Retrieval-Augmented Generation (RAG) solution that uses a Foundry IQ knowledge base. The solution includes a support agent and a policy agent.
You have an evaluation dataset that contains the following for each user query
* The expected source IDs
* Retrieved chunks in rank order
* The final agent response
You discover that an embedding model change and a custom analyzer change cause the failed traces shown in the following table.

You need to isolate the failing parts of the RAG pipeline.
Which evaluator should you use for each agent? To answer, drag the appropriate evaluators to the correct agents. Each evaluator may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content NOTE: Each correct selection is worth one point.

정답:

Explanation:
Support agent: Document Retrieval; Policy agent: Document Retrieval.
Both failures are retrieval failures and should therefore be diagnosed with the Document Retrieval evaluator.
The dataset includes expected source IDs and ranked retrieved chunks, which provides the retrieval ground truth required by that evaluator. For the support agent, the expected source appears too low in the ranking after unrelated chunks. For the policy agent, the required exception-table source is missing entirely.
Groundedness would evaluate whether the final response is supported by the context it actually received; it could therefore pass even when the retrieval stage failed to fetch a required source. Microsoft Foundry ' s Document Retrieval evaluator compares retrieved documents against labeled ground truth and reports ranking
/search metrics such as NDCG and Fidelity. The corrected answer is Document Retrieval for both agents. For operational use, the measurement should be captured in a repeatable dataset, trace, or automated gate so that the same criterion can be compared across versions. That is more useful than a one-off manual observation and makes regressions visible before they become production incidents.
Official Microsoft reference: Microsoft Foundry - RAG evaluators

You have a Microsoft Foundry multi-agent customer support solution that retrieves grounding data from a shared vector index. The indexed corpus contains product runbooks in Markdown and support articles in HTML Both document types use a consistent hierarchical markup.
You discover that current fixed-size token chunking creates chunks that cross section boundaries.
You need to recommend a chunking approach for the ingestion pipeline. The solution must preserve existing document structure boundaries and minimize custom chunking code.
What should you recommend?

• A.format-specific chunking with header splitters
• B.fixed-size chunking with token overlap
• C.recursive character chunking with structure-aware separators
• D.semantic chunking with topic-shift detection