Sample question 1 — Plan and Manage Azure AI Solutions
A defense contractor needs named entity recognition and key phrase extraction from the Azure AI Language service to run inside a classified facility where no outbound internet access exists. The workload will run on a Linux host that the team already operates, and the model data must stay on that host. Which approach should the architect recommend?
- A. Deploy the Language service container image on the host and call its local endpoint.
- B. Deploy a multi-service Azure AI resource and call it from the classified network.
- C. Deploy an Azure AI Search instance in a private virtual network with a firewall.
- D. Deploy a batch transcription job that uploads the documents when the host reconnects.
Answer: A. Option A is correct because the Language service container image runs the same text analytics models on the managed host and answers through a local endpoint with no internet dependency. Option B still needs outbound connectivity to the cloud endpoint, which the facility forbids. Option C is a search and enrichment platform rather than a text analytics engine. Option D defers the work until connectivity returns, which delays every result and moves the documents off the host.