AWS Certified Generative AI Developer - Professional (AIP-C01) (AIP-C01) AIP-C01 Questions 76–150 | Page 2/3

Multi-Selecthard

Which THREE factors should you consider when tuning your inference endpoint for cost-efficiency?

Select 3 answers

A.Auto Scaling policies based on custom metrics

B.Model quantization level

C.CloudTrail log retention period

D.Number of S3 buckets used

E.Optimal instance selection (e.g., inf2 vs p4d)

AnswersA, B, E

Proper scaling prevents over-provisioning.

Why this answer

Instance type, auto-scaling configuration, and model optimization (like quantization) are critical levers for cost efficiency.

Full explanation →

Multi-Selectmedium

Which THREE items must be configured when setting up a new Knowledge Base in Amazon Bedrock?

Select 3 answers

A.Embedding model selection.

B.Model evaluation dataset.

C.Vector database configuration.

D.Data source (e.g., S3).

E.External firewall rules.

AnswersA, C, D

Required for indexing.

Why this answer

Embedding model, data source, and vector database are the three mandatory components for a Knowledge Base.

Full explanation →

Multi-Selecthard

Which THREE strategies should be implemented to ensure responsible AI practices in a production workload?

Select 3 answers

A.Perform regular model evaluation for bias

B.Exclude all logging to protect privacy

C.Maintain comprehensive logs for auditability

D.Implement human-in-the-loop review for critical outputs

E.Allow users to train the model directly

AnswersA, C, D

Essential for fairness.

Why this answer

Bias testing, transparency/logging, and continuous monitoring are pillars of responsible AI.

Full explanation →

MCQmedium

You need to ensure that all data sent to an Amazon Bedrock model is encrypted in transit using private connectivity. What is the recommended service?

A.AWS WAF

B.AWS CloudFront

C.AWS PrivateLink (VPC Endpoints)

D.AWS Certificate Manager

AnswerC

PrivateLink provides private connectivity to Bedrock services within the VPC.

Why this answer

Interface VPC Endpoints (AWS PrivateLink) allow traffic to stay within the AWS network, avoiding the public internet.

Full explanation →

MCQmedium

You are configuring a Knowledge Base for Amazon Bedrock. You need to ingest data from an S3 bucket that contains sensitive financial reports. Which feature allows you to ensure the data is encrypted during the ingestion and retrieval process?

A.AWS Secrets Manager for S3 credentials

B.Amazon Bedrock Model invocation logs

C.AWS KMS with customer-managed keys (CMKs)

D.S3 Bucket policies only

AnswerC

Bedrock supports CMKs to encrypt the vector index and data stored in the knowledge base.

Why this answer

KMS integration is required for the knowledge base index to ensure data at rest is encrypted using customer-managed keys.

Full explanation →

MCQhard

You notice that your SageMaker model endpoint experiences cold starts. Which feature can mitigate this?

A.Increasing the instance count

B.Changing the container image

C.Using a smaller model

D.Enabling Provisioned Concurrency

AnswerD

Provisioned Concurrency pre-warms resources to avoid cold starts.

Why this answer

Provisioned Concurrency maintains a set number of initialized instances, eliminating cold starts for serverless inference.

Full explanation →

MCQhard

A company is using a third-party model via Bedrock and needs to ensure that the model provider does not use their data to train their base models. What must they do?

A.Disable all logging

B.Enable the 'Opt-out' setting in the Bedrock console

C.Understand that AWS does not use customer data to train their foundation models

D.Encrypt all data with a customer-managed key

AnswerC

This is a fundamental tenet of the AWS Bedrock service agreement.

Why this answer

Amazon Bedrock operates under the AWS service terms, which state that AWS does not use customer data to train foundation models, including third-party models, unless the customer specifically opts in.

Full explanation →

MCQeasy

Which Bedrock API parameter allows you to define a system-level role or persona for the model?

A.system

B.messages

C.top_p

D.temperature

AnswerA

System instructions are defined here.

Why this answer

The system message is used to set the persona and high-level instructions for the conversation.

Full explanation →

MCQeasy

What is the primary function of the Bedrock 'Model Invocation Logs'?

A.To throttle incoming requests.

B.To capture and store prompt/response payloads for auditing.

C.To perform real-time model retraining.

D.To reduce inference costs.

AnswerB

This is the exact purpose of invocation logging.

Why this answer

These logs are used for auditing and debugging by capturing the request and response payloads.

Full explanation →

Multi-Selectmedium

A company is setting up a governance framework for AI development. Which TWO of the following tools should they use to enforce safety and compliance in Amazon Bedrock?

Select 2 answers

A.AWS Config.

B.Amazon Bedrock Model Evaluation jobs.

C.Amazon Inspector.

D.AWS Systems Manager.

E.Amazon Bedrock Guardrails.

AnswersB, E

Allows for standardized safety and compliance benchmarking.

Why this answer

Guardrails and Model Evaluation jobs are the primary tools for enforcing safety and assessing compliance.

Full explanation →

MCQeasy

A developer wants to choose a foundation model in Amazon Bedrock that offers the best balance of reasoning capability and latency for a customer support chatbot. Which tool should the developer use to evaluate models?

A.Amazon CloudWatch Metrics for Bedrock.

B.AWS Glue DataBrew.

C.Bedrock Model Evaluation jobs.

D.Amazon SageMaker Model Monitor.

AnswerC

This service is specifically designed to evaluate and compare foundation models.

Why this answer

The Bedrock Model Evaluation service allows developers to compare model performance using automated and human-in-the-loop metrics.

Full explanation →

Multi-Selecthard

When designing for compliance, which THREE actions are part of the 'Shared Responsibility Model' for an AWS customer?

Select 3 answers

A.Maintaining the physical security of the data center.

B.Configuring encryption for the data at rest.

C.Managing access control through IAM policies.

D.Patching the underlying foundation model code.

E.Classifying data to determine if it is sensitive.

AnswersB, C, E

Customer responsibility.

Why this answer

Data classification, encryption configuration, and IAM management are the customer's responsibility.

Full explanation →

Multi-Selecthard

Which THREE items should be included in an AI application's security documentation for an audit?

Select 3 answers

A.Logging and monitoring policies

B.Encryption methods for training/RAG data

C.Public links to all internal datasets

D.User passwords for all employees

E.Description of safety filters (Guardrails)

AnswersA, B, E

Required for compliance.

Why this answer

Documentation should cover data handling, safety controls, and logging/audit policies.

Full explanation →

Multi-Selectmedium

When implementing a RAG-based application, which TWO of the following ensure document security?

Select 2 answers

A.Apply metadata filters to control document access

B.Use IAM-based connectors to data sources

C.Use a single shared user for all data source connections

D.Make all documents globally readable

E.Disable indexing of sensitive documents

AnswersA, B

Ensures document-level access control.

Why this answer

Metadata-based filtering and IAM-integrated connectors are the standard methods.

Full explanation →

Multi-Selecthard

Which THREE factors influence the choice of embedding model for a RAG solution?

Select 3 answers

A.Model's popularity

B.Supported languages

C.Maximum input token length

D.Model training cost

E.Embedding dimensionality

AnswersB, C, E

Must support the data language.

Why this answer

Language support, embedding dimensions, and maximum input token length are key factors.

Full explanation →

MCQeasy

Which API call in Amazon Bedrock is used to trigger a text generation request without a Knowledge Base?

A.InvokeModel

B.RetrieveAndGenerate

C.GetModelInvocation

D.CreateKnowledgeBase

AnswerA

This is the standard API for direct model inference.

Why this answer

InvokeModel is the primary API for sending a prompt directly to a foundation model.

Full explanation →

Multi-Selecthard

Which THREE of these are valid methods for monitoring a Bedrock-based application?

Select 3 answers

A.Direct access to model training logs

B.AWS CloudTrail for API calls

C.Application-level logging of requests

D.Manual binary inspection

E.Amazon CloudWatch Metrics

AnswersB, C, E

Audits access to Bedrock.

Why this answer

CloudWatch metrics, CloudTrail auditing, and logging the raw requests/responses are standard.

Full explanation →

MCQmedium

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

A.Amazon OpenSearch Serverless

B.Amazon DynamoDB

C.Amazon ElastiCache

D.Amazon RDS for PostgreSQL

AnswerA

OpenSearch Serverless is the primary native vector store for Bedrock.

Why this answer

Amazon OpenSearch Serverless is the native vector store integration for Bedrock Knowledge Bases.

Full explanation →

MCQeasy

Which of the following is a recommended practice when working with Bedrock APIs?

A.Use the root user for API calls.

B.Disable logging to save costs.

C.Hardcode IAM keys in the code.

D.Apply the principle of least privilege to IAM policies.

AnswerD

Standard security best practice.

Why this answer

Following the principle of least privilege in IAM is the most recommended security practice.

Full explanation →

Multi-Selecthard

Which THREE tasks are performed by a Bedrock Agent when an Action Group is invoked?

Select 3 answers

A.Mapping intent to an API definition

B.Interpreting user natural language intent

C.Automatically scaling the API endpoint

D.Invoking the defined Lambda function

E.Retraining the model on new data

AnswersA, B, D

Required for execution.

Why this answer

The agent parses the user intent, maps it to the API schema, and then calls the associated Lambda function.

Full explanation →

Multi-Selecthard

Which THREE configurations must be set correctly within a Knowledge Base 'Data Source' for Bedrock to ingest documents from Amazon S3?

Select 3 answers

A.IAM Role with S3 read permissions

B.Model ID of the generative model

C.Database write timeout

D.Chunking Strategy

E.S3 Bucket Path

AnswersA, D, E

Access control is required.

Why this answer

Ingestion requires specifying the S3 bucket, an IAM role for access, and the chunking strategy.

Full explanation →

Multi-Selecthard

A developer is troubleshooting an Amazon Bedrock Agent that is frequently hitting errors. Which TWO of the following steps should be taken to isolate the issue?

Select 2 answers

A.Restart the Bedrock service.

B.Increase the model's temperature to 1.0.

C.Check CloudWatch Logs for the associated Action Group Lambda functions.

D.Review the Agent's 'Trace' output to identify the failure point.

E.Re-train the underlying base model.

AnswersC, D

If the agent is failing during tool execution, the error is likely in the Lambda function code.

Why this answer

Examining the trace and checking the Lambda function logs are the standard diagnostic paths for Agent failures.

Full explanation →

Multi-Selectmedium

A team is preparing data for fine-tuning a custom model on Amazon Bedrock. Which TWO of the following are best practices for high-quality training datasets?

Select 2 answers

A.Include as much raw data as possible, even if noisy

B.Provide data in JSONL format with clear prompt-response pairs

C.Use automated AI tools to generate the training data from scratch

D.Convert all data into a CSV format

E.Ensure the dataset is representative of the production use case

AnswersB, E

Bedrock custom model training expects JSONL format for fine-tuning inputs.

Why this answer

Dataset preparation requires high-quality, relevant data and appropriate formatting (JSONL) for successful fine-tuning.

Full explanation →

MCQmedium

In the context of Bedrock Agents, what is the purpose of the 'Instruction' field in the agent configuration?

A.It defines the Knowledge Base search parameters

B.It serves as the system prompt to define the agent's behavior

C.It stores user credentials for the API

D.It contains the list of API definitions

AnswerB

The instruction field defines how the agent should act.

Why this answer

The instruction field serves as the system prompt that defines the agent's persona and objective.

Full explanation →

100

Multi-Selectmedium

A developer is creating a test suite for an LLM application. Which THREE of the following metrics are commonly used to evaluate RAG system performance?

Select 3 answers

A.Answer Relevance

B.Context Precision

C.Model Training Loss

D.Faithfulness

E.Inference Latency

AnswersA, B, D

Measures how relevant the answer is to the original query.

Why this answer

Faithfulness, answer relevance, and context precision are the key components of RAGAS evaluation.

Full explanation →

101

MCQhard

You are monitoring your Bedrock usage costs. You notice a spike in costs despite the number of requests remaining steady. What is the most likely cause?

A.The number of users has increased

B.The model has been upgraded to a newer version

C.The token count per request has increased

D.The regional data transfer cost has increased

AnswerC

Bedrock charges per token; longer inputs/outputs equal higher costs.

Why this answer

The size of the input tokens or the generated output tokens (which are billed) has significantly increased.

Full explanation →

102

MCQmedium

To ensure compliance with data residency requirements, you must restrict which AWS regions your Generative AI application can process data. What is the most effective way to enforce this?

A.Service Control Policies (SCPs)

B.VPC Endpoint policies

C.IAM policies for all users

D.CloudFront geographic restrictions

AnswerA

SCPs are the standard mechanism to enforce organizational boundaries across AWS accounts.

Why this answer

AWS Organizations Service Control Policies (SCPs) can be used to deny Bedrock service access in restricted regions.

Full explanation →

103

MCQeasy

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

A.AWS CloudTrail

B.Amazon CloudWatch

C.IAM Policy Simulator

D.AWS Config

AnswerC

Perfect for testing permissions.

Why this answer

IAM Policy Simulator allows you to verify if a specific role has the necessary permissions.

Full explanation →

104

Multi-Selecthard

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

Select 3 answers

A.Instance memory usage

B.Endpoint concurrency limits

C.Container health check status

D.IAM role permissions

E.Input JSON schema validation

AnswersA, B, C

Out-of-memory errors lead to 503s.

Why this answer

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

Full explanation →

105

MCQeasy

Which parameter in Bedrock inference controls the deterministic nature of the output?

A.Temperature

B.Top_p

C.Max_tokens

D.Stop_sequences

AnswerA

Temperature is the standard parameter for controlling randomness.

Why this answer

Temperature controls randomness; lower values make output more deterministic.

Full explanation →

106

MCQhard

An organization is integrating a custom model via Bedrock Custom Models. They need to ensure that the data used for fine-tuning remains within their VPC. Which configuration meets this requirement?

A.Using AWS Glue for data preparation only

B.Deploying the model in an on-premises data center

C.Configuring an interface VPC endpoint for Amazon Bedrock

D.Encrypting the dataset with AWS CloudHSM

AnswerC

VPC endpoints ensure traffic between the VPC and Bedrock stays within the AWS network.

Why this answer

AWS PrivateLink for Amazon Bedrock allows the traffic to never traverse the public internet, keeping the data flow contained within the VPC context.

Full explanation →

107

MCQmedium

You are designing a RAG pipeline. Which step in the ingestion process is responsible for converting unstructured text into a numerical representation?

A.Chunking

B.Embedding

C.Indexing

D.Parsing

AnswerB

Embedding creates the vector representation.

Why this answer

Embedding is the process of converting text into numerical vectors that can be searched for semantic similarity.

Full explanation →

108

MCQhard

A healthcare company requires that all generative AI outputs be filtered for PII (Personally Identifiable Information) before being sent to the end user. Which Amazon Bedrock feature effectively manages this compliance requirement?

A.Amazon Bedrock Guardrails

B.AWS CloudTrail logs

C.AWS WAF on the API Gateway

D.IAM Policy evaluation for Bedrock

AnswerA

Guardrails provides content filtering capabilities for both input prompts and model responses.

Why this answer

Guardrails for Amazon Bedrock allows developers to configure PII filtering, profanity filters, and denial topics to ensure compliance.

Full explanation →

109

MCQmedium

An application uses Amazon Bedrock InvokeModel to generate text. You notice that the output length is consistently truncated. Which parameter should you adjust to ensure the full response is captured?

A.Increase the max_tokens parameter

B.Increase the temperature parameter

C.Increase the top_p parameter

D.Decrease the stop_sequences parameter

AnswerA

Increasing max_tokens allows the model to generate more tokens.

Why this answer

The max_tokens (or max_gen_len) parameter limits the number of tokens in the generated output. Increasing this allows for longer responses.

Full explanation →

110

MCQeasy

A developer needs to share a Bedrock prompt template with another team securely. What is the recommended method?

A.Save the prompt as a text file in S3 with public access

B.Use Amazon Bedrock Prompt Management

C.Send the prompt in an email

D.Store the prompt in a Git repository without encryption

AnswerB

Prompt Management is the built-in tool for versioning and sharing prompts.

Why this answer

Amazon Bedrock Prompt Management allows for creating and sharing versioned prompt templates across accounts or teams securely.

Full explanation →

111

Multi-Selectmedium

Which TWO of the following are valid ways to improve the relevance of retrieved documents in a RAG system?

Select 2 answers

A.Implement metadata filtering to narrow the search space.

B.Apply a custom chunking strategy suited to the document format.

C.Use a larger, more generic embedding model.

D.Increase the number of retrieved chunks to 1000.

E.Delete all indices and recreate them every hour.

AnswersA, B

Filters improve precision.

Why this answer

Metadata filtering and choosing an optimal chunking strategy are key levers to improve retrieval accuracy.

Full explanation →

112

MCQeasy

When invoking a Bedrock model via the API, which HTTP status code indicates that the request was processed, but the model has reached its rate limit?

A.500 Internal Server Error

B.400 Bad Request

C.429 Too Many Requests

D.200 OK

AnswerC

429 means rate limit exceeded.

Why this answer

429 Too Many Requests is the standard status code for rate limiting.

Full explanation →

113

Multi-Selecthard

Which THREE factors influence the cost of running an Amazon Bedrock Knowledge Base?

Select 3 answers

A.The amount of data stored in the vector database.

B.The number of tokens generated by the model during the 'RetrieveAndGenerate' process.

C.The number of concurrent IAM users.

D.The number of tokens processed by the embedding model.

E.The frequency of S3 bucket scans.

AnswersA, B, D

Storage volume is billed.

Why this answer

Costs are driven by the embedding model usage, vector storage, and the generation model usage during retrieval.

Full explanation →

114

MCQmedium

A business wants to restrict the length of generated outputs to prevent high costs. How can this be done?

A.Set maxTokenCount in the inference configuration

B.Use Lambda to truncate text

C.Disable streaming

D.Use Guardrails to block long responses

AnswerA

The maxTokenCount parameter in the API request directly limits the response size.

Why this answer

In the inference parameters of the Bedrock API (or the Bedrock InvokeModel request), you can set 'maxTokenCount' to limit the response length.

Full explanation →

115

MCQhard

An application deployed on AWS Lambda is receiving 429 Too Many Requests errors when calling Amazon Bedrock. The developer has verified that account-level quotas have not been reached. What is the most likely cause?

A.The VPC endpoint policy is blocking traffic to the Bedrock runtime service.

B.The application is exceeding the tokens-per-minute (TPM) limit for the specific model in that AWS Region.

C.The Lambda function timeout is set too low for the model response time.

D.The IAM role is missing the bedrock:InvokeModelWithResponseStream permission.

AnswerB

Each model has specific throughput limits that, when exceeded, trigger 429 errors.

Why this answer

Model-specific throughput limits or Provisioned Throughput quotas are distinct from account-level service quotas.

Full explanation →

116

Multi-Selecthard

When dealing with PII in Generative AI, which THREE practices are recommended?

Select 3 answers

A.Store PII in plain text logs for debugging

B.Implement access policies to restrict PII visibility

C.Share PII freely across all AWS accounts

D.Ensure PII is encrypted at rest using KMS

E.Use Bedrock Guardrails PII redaction

AnswersB, D, E

Limits exposure.

Why this answer

Redaction, encryption, and limiting access are standard PII practices.

Full explanation →

117

MCQmedium

What is the primary role of an 'Embeddings Model' in a RAG-based Generative AI application?

A.To generate the final natural language answer

B.To convert input text into high-dimensional numerical vectors

C.To store document chunks for long-term retrieval

D.To parse PDF files into plain text

AnswerB

This is the definition of embedding.

Why this answer

Embeddings models translate text into vector space, enabling semantic retrieval.

Full explanation →

118

MCQhard

You are attempting to use a Bedrock Agent to query an internal database, but the agent constantly fails to understand the schema. What is the best way to improve the agent's performance?

A.Improve the OpenAPI schema documentation

B.Increase the agent's timeout

C.Use a Knowledge Base to store the schema

D.Upgrade the agent model

AnswerA

Better descriptions directly help the agent navigate the schema.

Why this answer

Providing a clear, descriptive OpenAPI schema with high-quality descriptions for each field is critical for agent understanding.

Full explanation →

119

MCQeasy

Which AWS tool is used to monitor the performance and latency of an application calling Bedrock?

A.Amazon Inspector

B.AWS CloudTrail

C.Amazon CloudWatch

D.AWS Config

AnswerC

CloudWatch is the standard monitoring service for AWS infrastructure and services.

Why this answer

Amazon CloudWatch provides metrics for request latency, invocation count, and errors for Bedrock models.

Full explanation →

120

MCQeasy

When using the Bedrock Converse API, which object allows you to maintain context across multiple turns of a conversation?

A.inferenceConfig

B.additionalModelRequestFields

C.messages

D.system

AnswerC

Messages holds the sequence of turns.

Why this answer

The messages array contains the chat history (user and assistant turns) needed to maintain state.

Full explanation →

121

MCQeasy

Which AWS service is used to monitor and log the prompts sent to an Amazon Bedrock model for auditing purposes?

A.AWS Config

B.AWS CloudTrail

C.Amazon Macie

D.Amazon CloudWatch Logs

AnswerB

CloudTrail captures the API request details including the prompt content.

Why this answer

AWS CloudTrail logs all API calls made to Amazon Bedrock, providing the necessary audit trail for prompt usage.

Full explanation →

122

Multi-Selectmedium

Which THREE techniques are commonly used to mitigate hallucinations in RAG applications?

Select 3 answers

A.Adding 'I don't know' as a fallback response

B.Using the longest possible output token limit

C.Requiring the model to provide citations

D.Instructing the model to only answer based on provided context

E.Increasing the inference temperature

AnswersA, C, D

Prevents guessing.

Why this answer

System prompts, retrieval-augmented generation (RAG) itself, and strict source citation reduce hallucinations.

Full explanation →

123

MCQhard

A company wants to provide users with citations for the AI-generated answers in their RAG system. How is this achieved in Amazon Bedrock?

A.Manually write code to compare generation results with source documents.

B.Use the RetrieveAndGenerate API which includes source attribution.

C.Configure the Prompt Template to force the model to output document IDs.

D.Enable the 'Citations' feature in the Model Evaluation console.

AnswerB

The managed service automatically tracks source documents and returns citation references.

Why this answer

When using Knowledge Bases for Amazon Bedrock, the response includes source attribution and citations by default when the 'RetrieveAndGenerate' API is used.

Full explanation →

124

MCQmedium

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

A.It caches the model responses

B.It automatically formats the output as JSON

C.It provides consistent instructions and persona for the model

D.It overrides the model's training data

AnswerC

This is the primary use case for system messages.

Why this answer

System prompts allow you to define the model's behavior, tone, and constraints before the conversation begins.

Full explanation →

125

MCQeasy

What is the role of an AWS Service Control Policy (SCP) in an AI governance framework?

A.To store model weights securely

B.To automatically optimize model latency

C.To monitor user chat history

D.To enforce organizational compliance and restrict service access

AnswerD

SCPs are for organizational-level permission enforcement.

Why this answer

SCPs provide centralized guardrails to define the maximum permissions that can be granted to IAM roles across all accounts in an organization.

Full explanation →

126

MCQmedium

What is the recommended approach to handle PII in your RAG Knowledge Base to comply with privacy regulations?

A.Redact PII before ingesting documents into the Knowledge Base

B.Filter it during query time only

C.Encrypt the Knowledge Base with a master key

D.Use a specialized model that doesn't 'read' PII

AnswerA

Pre-ingestion redaction ensures PII never enters the vector store.

Why this answer

You should redact or mask PII at the source before ingestion into the Knowledge Base.

Full explanation →

127

MCQeasy

Which service should be used to manage user access to Amazon Bedrock models?

A.Amazon Route 53.

B.Amazon SageMaker.

C.AWS Glue.

D.AWS IAM.

AnswerD

IAM provides the granular control needed for Bedrock model access.

Why this answer

AWS IAM (Identity and Access Management) is the service used to control access to AWS resources, including Bedrock APIs.

Full explanation →

128

MCQhard

Which configuration in an Amazon Bedrock Knowledge Base is essential for updating the vector index whenever the source files in S3 change?

A.Configure Data Source Syncing.

B.Use AWS Config to monitor bucket changes.

C.Set the knowledge base to 'Auto-Refresh' mode.

D.Enable S3 Versioning.

AnswerA

The 'Sync' job updates the vector index to reflect changes in the source bucket.

Why this answer

Syncing is triggered via the 'Sync' operation, but for automation, S3 Event Notifications coupled with a Lambda function or the Bedrock Sync API is required.

Full explanation →

129

MCQmedium

A developer is troubleshooting a slow retrieval process in a RAG system. What is the most likely cause related to the vector store?

A.The vector index is missing an appropriate HNSW index or is not properly partitioned.

B.The embedding model is too small.

C.The source documents are too short.

D.The Bedrock service is overloaded.

AnswerA

Lack of proper indexing mechanisms causes linear search times in high-volume vector stores.

Why this answer

If the vector store lacks an index or the index is not partitioned correctly, search latency increases significantly.

Full explanation →

130

MCQhard

How can a developer ensure that the retrieval step in a RAG application focuses only on the most relevant documents when the knowledge base contains millions of files?

A.Implement metadata filtering in the query.

B.Increase the vector dimension size.

C.Use a larger model for embedding.

D.Reduce the chunk size.

AnswerA

Filtering by metadata (e.g., date, category) significantly improves relevance by excluding irrelevant partitions.

Why this answer

Using metadata filtering allows the system to narrow down the search space before performing vector similarity searches, increasing accuracy.

Full explanation →

131

MCQhard

A developer is optimizing a RAG solution for a legal firm. The documents are highly dense and technical. Which chunking strategy in Amazon Bedrock Knowledge Bases provides the best balance of retrieval accuracy?

A.Hierarchical chunking.

B.Semantic chunking based on paragraph breaks.

C.Fixed-size chunking with 100 token overlap.

D.Default chunking strategy.

AnswerA

This approach preserves relationships between document segments, which is critical for technical and legal density.

Why this answer

Hierarchical chunking is designed for complex documents by creating parent and child chunks, which helps maintain context during retrieval.

Full explanation →

132

MCQhard

An organization is concerned about 'Model Hallucinations' and needs a way to ground model outputs in trusted company documents. Which service supports this?

A.Amazon Bedrock Knowledge Bases

B.Amazon SageMaker Feature Store

C.AWS Glue Data Catalog

D.Amazon DynamoDB

AnswerA

Knowledge Bases automate the RAG workflow to ground responses in enterprise data.

Why this answer

Amazon Bedrock Knowledge Bases provide RAG (Retrieval-Augmented Generation), which grounds model responses in provided documents.

Full explanation →

133

MCQeasy

A company is using Amazon Q Business. How can they restrict access to specific data sources for different groups of users?

A.Create separate Amazon Q applications

B.Manually delete documents for specific users

C.Configure ACLs in the data source connector

D.Apply IAM policies to the user's login

AnswerC

Amazon Q honors the document-level security and ACLs from integrated data sources.

Why this answer

Amazon Q Business uses access control lists (ACLs) synced from the data source to ensure users only retrieve information they have permissions to access.

Full explanation →

134

MCQeasy

Which SageMaker feature allows you to observe model performance drifts over time?

A.SageMaker Pipelines

B.SageMaker Data Wrangler

C.SageMaker Clarify

D.SageMaker Model Monitor

AnswerD

Model Monitor is designed to track quality and drift.

Why this answer

SageMaker Model Monitor automatically detects data drift, concept drift, and model quality issues.

Full explanation →

135

MCQeasy

To track the resource usage of a SageMaker training job, you should look at which CloudWatch metric?

A.GPUUtilization

B.RequestCount

C.ModelLatency

D.Invocations

AnswerA

This metric tracks the hardware performance of the training node.

Why this answer

SageMaker training jobs publish GPUUtilization, CPUUtilization, and MemoryUtilization to CloudWatch under the /aws/sagemaker/TrainingJobs namespace.

Full explanation →

136

Multi-Selectmedium

Which TWO of the following actions can help mitigate the risk of 'prompt injection' in a Generative AI application?

Select 2 answers

A.Disable the chat interface

B.Set the temperature to 0.0

C.Use the most powerful foundation model available

D.Sanitize user input before passing it to the model

E.Use Bedrock Guardrails for input filtering

AnswersD, E

Input sanitization is a core security practice.

Why this answer

Input validation and using Guardrails are the primary defenses against prompt injection.

Full explanation →

137

Multi-Selecthard

You are auditing an AI application for security risks. Which THREE of the following are potential threats that should be tested against the model?

Select 3 answers

A.Toxic or biased output generation.

B.Insecure API endpoint access.

C.Excessive API latency.

D.Prompt injection attacks.

E.PII data leakage in the response.

AnswersA, D, E

Ensuring the model adheres to safety policies regarding tone and content.

Why this answer

Prompt injection, PII leakage, and model output toxicity are key security concerns for LLM applications.

Full explanation →

138

MCQmedium

You are building a chat application and notice sensitive internal data is leaking in responses. You have enabled Guardrails. What else should be checked?

A.Change the foundation model provider

B.Increase the temperature setting

C.Check if the model is fine-tuned on public data

D.Review the system prompt and the Knowledge Base document retrieval filters

AnswerD

If leakage occurs, either the RAG retrieval is pulling too much data or the system prompt isn't correctly enforcing safety rules.

Why this answer

The system prompt might be overly permissive or failing to reinforce the boundaries defined in the Guardrail.

Full explanation →

139

Multi-Selectmedium

Which TWO of the following are supported as source data for Amazon Bedrock Knowledge Bases?

Select 2 answers

A.Atlassian Confluence.

B.Email server logs.

C.Social media live streams.

D.Local hard drive files.

E.Amazon S3.

AnswersA, E

Supported natively.

Why this answer

S3 is the primary supported source, and Confluence is also natively supported as a data source.

Full explanation →

140

MCQmedium

An application using Amazon Bedrock is experiencing high latency. You need to monitor the 'ModelInvocationLatency' metric to identify performance bottlenecks. Which AWS service provides the primary integration for this telemetry?

A.AWS CloudTrail

B.AWS X-Ray

C.Amazon CloudWatch

D.Amazon OpenSearch Service

AnswerC

Bedrock publishes standard operational metrics directly to CloudWatch.

Why this answer

Amazon CloudWatch is the service that collects and monitors metrics from Amazon Bedrock, including model invocation latency.

Full explanation →

141

MCQhard

You are implementing a multi-agent system. Which method is most effective for ensuring the agents do not loop infinitely when coordinating tasks?

A.Caching agent responses

B.Increasing the model temperature

C.Using a smaller model

D.Implementing a maximum iteration count

AnswerD

Hard limits prevent infinite loops.

Why this answer

Defining a clear termination condition or a maximum iteration count prevents infinite agent loops.

Full explanation →

142

Multi-Selecthard

An enterprise is planning to implement generative AI with strict compliance requirements. Which THREE practices should the developer include to ensure data privacy and regulatory compliance?

Select 3 answers

A.Enable Amazon Bedrock Guardrails to filter PII

B.Use AWS KMS for encrypting the knowledge base index

C.Use IAM roles to enforce the principle of least privilege

D.Disable all CloudWatch logging to prevent data leakage

E.Store all prompts in plain text in S3 for auditing

AnswersA, B, C

Guardrails are specifically designed to meet compliance standards regarding input/output content.

Why this answer

Proper compliance involves encryption, access control (IAM), and the use of Guardrails for content moderation.

Full explanation →

143

Multi-Selecteasy

Which THREE of the following are valid ways to monitor Amazon Bedrock model usage?

Select 3 answers

A.AWS CloudTrail logs.

B.Amazon CloudWatch metrics.

C.AWS X-Ray service map.

D.AWS Billing and Cost Management reports.

E.Amazon S3 bucket logging.

AnswersA, B, D

Tracks API calls and administrative actions.

Why this answer

CloudWatch metrics, AWS CloudTrail logs, and Usage billing reports are the primary ways to monitor usage.

Full explanation →

144

MCQhard

You are implementing a Bedrock Guardrail. You need to ensure the model filters out specific PII (Personally Identifiable Information). Which capability should be configured?

A.Content Filter

B.Denied Topics

C.Sensitive Information Filter

D.Word Policy

AnswerC

This is the specific capability for PII.

Why this answer

Sensitive Information Filters within Guardrails allow the blocking of PII such as emails and social security numbers.

Full explanation →

145

MCQhard

You are troubleshooting a SageMaker inference endpoint and need to see the container logs. Which service stores these logs?

A.Amazon CloudWatch Logs

B.Amazon S3

C.Amazon DynamoDB

D.AWS CloudTrail

AnswerA

This is the default destination for container runtime logs.

Why this answer

SageMaker writes container logs (stdout/stderr) to CloudWatch Logs under the /aws/sagemaker/Endpoints/ path.

Full explanation →

146

MCQmedium

A developer is preprocessing a large corpus of PDF documents for RAG. To ensure the highest accuracy for multi-modal retrieval tasks, which data strategy is most effective?

A.Converting all text to images before embedding

B.Applying semantic chunking and adding metadata tags for filtering

C.Using a fixed-size character count of 500 without overlap

D.Concatenating all chunks into one large vector

AnswerB

Semantic chunking maintains logical document units, and metadata allows for surgical retrieval.

Why this answer

Chunking and metadata enrichment allow the retriever to locate specific sections rather than retrieving entire documents, improving precision.

Full explanation →

147

MCQeasy

Which AWS service is used to centralize and visualize costs for your various SageMaker projects?

A.CloudWatch Dashboards

B.AWS Cost Explorer

C.SageMaker Studio Dashboard

D.AWS Trusted Advisor

AnswerB

Cost Explorer is the primary tool for cost visualization.

Why this answer

AWS Cost Explorer allows you to view, analyze, and visualize your AWS spending, including SageMaker usage.

Full explanation →

148

MCQeasy

Which service should be used to provide secure identity and access management for an application accessing Bedrock?

A.AWS Secrets Manager

B.AWS Directory Service

C.Amazon Cognito

D.AWS IAM

AnswerD

IAM is the standard for service-to-service and identity-based access control.

Why this answer

AWS IAM is the core service for managing identities and permissions for all AWS services, including Bedrock.

Full explanation →

149

Multi-Selectmedium

A company is designing a RAG architecture. Which TWO of the following steps are required to successfully link an Amazon S3 source to an Amazon Bedrock Knowledge Base?

Select 2 answers

A.Configure an IAM role with S3 read access for Bedrock

B.Choose an embedding model available in Bedrock

C.Create a Glue Data Catalog for the S3 bucket

D.Enable AWS WAF on the S3 bucket

E.Create a Lambda function to split PDFs manually

AnswersA, B

Bedrock requires permissions to read the source documents from S3.

Why this answer

The process requires selecting the data source (S3) and specifying the embedding model to convert text to vectors.

Full explanation →

150

MCQmedium

A developer needs to measure the quality of a RAG application's responses. Which tool within Amazon Bedrock facilitates this?

A.Amazon CloudWatch Metrics.

B.Amazon Bedrock Model Evaluation.

C.AWS Glue Quality Rules.

D.Amazon SageMaker Model Monitor.

AnswerB

This service specifically supports evaluating model outputs for accuracy and relevance.

Why this answer

Amazon Bedrock Model Evaluation allows users to evaluate their models or RAG pipelines against custom or built-in datasets.

Full explanation →

AWS Certified Generative AI Developer - Professional (AIP-C01) (AIP-C01) (AIP-C01) — Questions 76–150