Task 1.5: Design Retrieval Mechanisms for FM Augmentation. AIP-C01 Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A company is building a search system for a large e-commerce catalog. They have a mix of structured product attributes (e.g., brand, price) and unstructured descriptions. They want to support both faceted search (e.g., filter by brand) and semantic search (e.g., 'comfortable running shoes'). They are using Amazon OpenSearch Service. What is the best architecture?

   • AUse Amazon Bedrock Knowledge Bases with OpenSearch Serverless, which automatically handles hybrid search.
   • BUse a single OpenSearch index with both keyword fields for facets and a vector field for semantic search, and use a hybrid query.
   • CUse a Lambda function to perform keyword search and a separate Lambda for vector search, then combine results.
   • DUse two separate indexes: one for keyword search and one for vector search, and merge results in the application.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   A developer is building a retrieval system where a user query is transformed into multiple sub-queries to improve search coverage. Which technique is being used?

   • AQuery decomposition
   • BQuery expansion
   • CQuery reranking
   • DQuery embedding
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   A developer needs to generate embeddings for a domain-specific corpus of legal documents. The developer wants to evaluate the performance of different embedding models available in Amazon Bedrock. Which factor is most important to consider when comparing embedding models for this use case?

   • AThe maximum number of tokens the model can process in a single request.
   • BThe availability of the model in the AWS region where the application is deployed.
   • CThe dimensionality of the output vectors and the model's ability to capture domain-specific semantics.
   • DThe cost per million tokens for generating embeddings.
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   A developer is building a system that allows a foundation model to access external data sources through a standardized protocol. Which protocol is designed to connect AI models to various data sources and tools?

   • AREST API
   • BWebSocket
   • CGraphQL
   • DModel Context Protocol (MCP)
   Reveal Solution Discussion Select an answer first
6. 15expert · hard

   A company is building a multi-agent system where multiple FMs need to access the same vector search index. They want to use a standardized protocol to integrate the vector search with the FMs. They are considering using Model Context Protocol (MCP). What is the best way to implement this?

   • AUse an MCP client in each agent to connect to an MCP server that wraps the vector search API.
   • BUse a Lambda function as a tool for each agent via function calling.
   • CUse a direct API call from each agent to the vector search API.
   • DUse a Step Functions workflow to orchestrate the retrieval for each agent.
   Reveal Solution Discussion Select an answer first