Design and Implement Models and Embeddings DP-800 Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   A database solution needs to return AI-generated responses in a consistent JSON format for downstream applications. Which evaluation factor is most important when selecting an external AI model?

   • AModel size
   • BMultimodal capabilities
   • CStructured output support
   • DMultilanguage support
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A research organization stores scientific papers in a SQL database. They want to generate embeddings for semantic search. The papers contain complex tables and figures. The embedding model has a token limit of 2,048. The team needs to decide how to chunk the papers to maximize search relevance. They are considering (A) chunking by sections with overlap, (B) chunking by fixed token count, (C) chunking by paragraphs, and (D) embedding the entire paper. Which approach is most effective?

   • AChunk by fixed token count
   • BChunk by sections with overlap
   • CChunk by paragraphs
   • DEmbed the entire paper
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A news website stores articles in Azure SQL Database. New articles are inserted continuously, and existing articles are occasionally edited. The team needs to maintain embeddings for a search feature. They want to use Change Tracking to identify changed rows. Which additional component is required to generate embeddings for only the changed rows?

   • AAn Azure Logic App that runs on a schedule and re-embeds the entire table.
   • BA SQL Server Agent job that regenerates embeddings for all rows nightly.
   • CA database trigger that directly updates the embedding column.
   • DAn Azure Function that polls the change tracking table and calls the embedding API for each changed row.
   Reveal Solution Discussion Select an answer first
5. 14expert · hard

   A company maintains embeddings for a large product catalog in Azure SQL Database. They need to update embeddings for changed rows within 5 minutes of a change. They are considering Change Tracking with an Azure Function, Azure Logic Apps, and CDC. The team has a preference for minimal custom code and low operational overhead. Which solution should they choose?

   • AA scheduled Azure Data Factory pipeline every 10 minutes
   • BCDC with a custom polling service
   • CChange Tracking with an Azure Function using the SQL trigger binding
   • DAzure Logic Apps with a SQL trigger
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   In Azure SQL Database, which T-SQL statement is used to create an external AI model that references a model hosted in Azure Machine Learning?

   • ACREATE AI MODEL
   • BCREATE EXTERNAL TABLE
   • CCREATE EXTERNAL MODEL
   • DCREATE MODEL
   Reveal Solution Discussion Select an answer first