AIP-C01 Implementation And Integration — Timed 20 Questions

Question 1

You are optimizing a prompt for a Large Language Model to improve reasoning capabilities. Which TWO techniques are most effective for improving zero-shot performance?

Accepted Answer

Persona adoption. Chain-of-thought and persona adoption are industry-standard techniques for improving zero-shot reasoning.

Answer

Embedding the prompt in a Knowledge Base

Answer

Few-shot prompting

Answer

Increasing temperature to 1.0

Question 2

Which Bedrock API operation is the recommended unified interface to stream responses from multiple foundation models?

Accepted Answer

Converse. The Converse API is designed as a unified, consistent API for interacting with various models in Bedrock.

Answer

GetFoundationModel

Answer

ListFoundationModels

Answer

InvokeModel

Question 3

When troubleshooting a SageMaker endpoint that consistently returns 503 Service Unavailable errors, which THREE areas should you investigate?

Accepted Answer

Instance memory usage. 503s typically point to capacity limits, health check failures, or resource exhaustion (memory/GPU).

Answer

IAM role permissions

Answer

Input JSON schema validation

Question 4

You are using SageMaker JumpStart to deploy a text-generation model. Which THREE deployment configurations are mandatory when setting up the Endpoint Configuration?

Accepted Answer

Instance Type. Instance type, variant name, and model name are essential to define the endpoint configuration.

Answer

Spot Instance Configuration

Answer

Auto-scaling policy

Question 5

You are building an agentic application. Which THREE of the following are valid ways for an agent to perform an action?

Accepted Answer

Invoking a Lambda function. Option B is correct because invoking a Lambda function is a canonical way for an agent to execute an action through a defined tool/function interface, letting the agent trigger server-side compute with specific parameters. Option C is correct because querying a Knowledge Base is a valid agent action that retrieves relevant information (typically via retrieval APIs such as Amazon Bedrock Knowledge Bases Retrieve/RetrieveAndGenerate) to ground the agent's next step. Option E is correct because returning a direct response to the user is itself a valid action the agent can take, completing the turn with a natural-language answer when no external tool call is needed. Option A does not belong because directly modifying a database table without an API bypasses controlled, auditable tool interfaces and is not a supported agent action pattern. Option D does not belong because running a local Docker container is an infrastructure/runtime operation, not an agent action exposed through the agent's tool/action model.

Answer

Directly modifying a database table without an API

Answer

Running a local Docker container

Question 6

Which THREE of the following are valid ways to configure inference parameters in Bedrock?

Accepted Answer

Passing parameters in the InvokeModel request body. Parameters can be passed directly in InvokeModel, managed via Prompt Management, or set as default configurations.

Answer

Directly editing the Bedrock service binary

Answer

Configuring them inside the model's training script

Question 7

What is the primary benefit of using a 'System Prompt' in Bedrock?

Accepted Answer

It provides consistent instructions and persona for the model. A System Prompt in Amazon Bedrock sets the persistent context, persona, and behavioral instructions for the model across a conversation, ensuring consistent tone, format, and constraints. It is prepended to the user's input and influences all subsequent generations without being repeated by the user each turn. This is the primary benefit: consistent instructions and persona.

Answer

It caches the model responses

Answer

It automatically formats the output as JSON

Answer

It overrides the model's training data

Question 8

Which Bedrock API allows for the asynchronous ingestion of data into a Knowledge Base?

Accepted Answer

StartIngestionJob. StartIngestionJob is the Bedrock Agents API that asynchronously begins ingesting data from a configured data source into a Knowledge Base. It returns an ingestionJobId that can be polled with GetIngestionJob to track progress, making it the correct API for loading or refreshing documents.

Answer

InvokeModel

Answer

ListKnowledgeBases

Answer

Retrieve

Question 9

You are deploying a custom model via SageMaker. Which THREE items must be present in the model package (.tar.gz) for successful deployment?

Accepted Answer

requirements.txt. Custom models require the model artifacts, inference code, and the requirements.txt file.

Answer

SSH private key

Answer

Training dataset

Question 10

You want to automate the ingestion of new PDF files added to an S3 bucket into a Bedrock Knowledge Base. What is the most efficient design?

Accepted Answer

Use an S3 Event Notification to trigger a Lambda function that calls StartIngestionJob. Using an EventBridge rule to trigger a Lambda function that calls the Bedrock Knowledge Base sync API is the standard event-driven approach.

Answer

Schedule a daily batch cron job

Answer

Use AWS Glue to ETL the data directly to the vector store

Answer

Manually update the Knowledge Base via the console

Question 11

Which THREE configuration parameters are standard in a SageMaker JumpStart inference request?

Accepted Answer

inputs. Standard inference payloads typically require inputs, temperature, and max_tokens.

Answer

model_owner

Answer

S3_output_path

Question 12

Which service should you check first to troubleshoot IAM permission issues when accessing Bedrock?

Accepted Answer

IAM Policy Simulator. The IAM Policy Simulator is the first tool to check because it lets you test the exact API action (e.g., bedrock:InvokeModel) against a specific principal, resource, and condition context to see whether the policy allows or denies it. It evaluates all applicable identity-based policies, resource policies, permissions boundaries, and SCPs, showing which statement caused the deny. This directly answers 'why is this IAM permission failing?' without needing to reproduce the call.

Answer

AWS CloudTrail

Answer

Amazon CloudWatch

Answer

AWS Config

Question 13

You are building a RAG application. Which metadata field is most useful for filtering results during retrieval?

Accepted Answer

Creation date. Metadata filtering allows you to restrict retrieval to specific document types or categories.

Answer

Vector distance

Answer

Token count

Answer

Model ID

Question 14

Which parameter in Bedrock inference controls the diversity of the generated tokens?

Accepted Answer

Top_p. Top_p (also called nucleus sampling) limits the token choices to the top percentage of probability mass, affecting diversity.

Answer

Max_tokens

Answer

Temperature

Answer

Top_k

Question 15

A company is deploying a generative AI chatbot on Amazon Bedrock that uses the Converse API. The security team requires that all user prompts and model responses be logged for auditing, but the logs must not contain any personally identifiable information (PII). The application already uses Amazon CloudWatch Logs. Which solution meets these requirements with the LEAST operational overhead?

Accepted Answer

Create an Amazon Bedrock guardrail with sensitive information filters for PII, attach it to the model invocation, and enable model invocation logging to CloudWatch Logs. The guardrail will redact PII before it is logged.. A Bedrock guardrail with sensitive information filters can automatically detect and redact PII in prompts and responses. When model invocation logging is enabled, the logged data reflects the redacted content, satisfying the audit requirement without exposing PII. This managed feature avoids custom code and reduces operational burden compared to building a redaction pipeline or relying on CloudTrail, which does not capture payloads.

Answer

Use AWS CloudTrail to log all Bedrock API calls, and configure CloudTrail to include request and response bodies. CloudTrail will automatically mask PII fields.

Answer

Implement a custom Lambda function that intercepts every Converse API call, redacts PII using Amazon Comprehend, and then writes the sanitized logs to CloudWatch Logs.

Answer

Enable model invocation logging in Amazon Bedrock and configure it to send logs to CloudWatch Logs, then use a CloudWatch Logs subscription filter with a Lambda function to redact PII before storing the logs.

Question 16

A developer is building a retrieval-augmented generation (RAG) application using Amazon Bedrock Knowledge Bases. The knowledge base is backed by an Amazon OpenSearch Serverless vector store. Users report that responses sometimes include outdated information even though the source documents have been updated in Amazon S3. Which TWO actions should the developer take to ensure the knowledge base reflects the latest document versions? (Choose two.)

Accepted Answer

Configure the knowledge base to use incremental sync and schedule it to run automatically after S3 updates.. To keep a Bedrock knowledge base current, you need to run ingestion jobs that detect and apply changes. Incremental sync updates only modified documents, reducing cost and time. Automating the sync with EventBridge and Lambda when S3 objects change ensures timely updates without manual intervention. Together, these actions maintain accurate and up-to-date responses.

Answer

Use Amazon Bedrock's automatic synchronization feature, which continuously monitors S3 and updates the knowledge base in real time.

Answer

Manually trigger a full sync of the knowledge base after each document update.

Answer

Enable versioning on the S3 bucket and configure the knowledge base to use the latest object version.

Question 17

When using the Amazon Bedrock Converse API, which parameter enables the model to perform tool use by providing a definition of available functions?

Accepted Answer

toolConfig. The tools parameter in the Converse API allows you to define a list of tools for the model to use.

Answer

inferenceConfig

Answer

system

Answer

additionalModelRequestFields

Question 18

You are deploying a model from SageMaker JumpStart. You need to ensure that the model inference container has access to private VPC resources. What must you configure?

Accepted Answer

Configure Subnets and Security Groups in the Model creation. You must configure the VPC settings (Subnets and Security Groups) in the SageMaker Model configuration to allow network access.

Answer

Set the IAM role to 'AdministratorAccess'

Answer

Update the JumpStart model Dockerfile

Answer

Enable AWS PrivateLink on the S3 bucket

Question 19

You are designing an Agentic AI solution using Bedrock Agents. Which THREE components must you configure to enable the agent to interact with an external API?

Accepted Answer

Action Group. Option B (Action Group) is correct because an action group is the Bedrock Agents construct that defines the set of tasks the agent can invoke against external APIs, binding the agent's reasoning to concrete operations. Option C (OpenAPI Schema) is correct because the action group requires an OpenAPI schema (API schema) that describes the external API's endpoints, parameters, and request/response structures so the agent knows how to call it. Option D (Lambda Function) is correct because the action group must be backed by an executor — typically an AWS Lambda function — that receives the agent's invocation and performs the actual API call to the external service. Option A (Knowledge Base) is not required here because knowledge bases are for retrieval-augmented generation over documents/data, not for invoking external APIs. Option E (Guardrail) is not required because guardrails enforce content safety and policy filtering, which is unrelated to enabling API interaction.

Answer

Knowledge Base

Answer

Guardrail

Question 20

When using Bedrock Knowledge Bases, which vector database is integrated natively to support managed RAG workflows?

Accepted Answer

Amazon OpenSearch Serverless. Amazon Bedrock Knowledge Bases natively integrates with Amazon OpenSearch Serverless as its managed vector store, allowing RAG workflows to be set up without provisioning or managing vector database infrastructure. OpenSearch Serverless provides the vector search collection type that Bedrock uses to store and retrieve embeddings during retrieval-augmented generation. This native integration means customers can point Bedrock at an OpenSearch Serverless vector index and have chunking, embedding, and retrieval handled automatically.

Answer

Amazon DynamoDB

Answer

Amazon ElastiCache

Answer

Amazon RDS for PostgreSQL