AIP-C01 Practice Test 13 — 15 Questions

Question 1

You are designing a secure Generative AI architecture on AWS. Which TWO of the following are recommended practices for securing the data ingestion pipeline for RAG?

Accepted Answer

Enforce TLS 1.2+ for all data movement. Encrypting data in transit (TLS) and at rest (KMS) are fundamental for data pipeline security.

Answer

Disable all logging in the pipeline to prevent data exposure

Answer

Store API keys in plain text within the Lambda code

Answer

Store all documents in a public S3 bucket for ease of access

Question 2

A developer is using Amazon Bedrock to build a customer support chatbot. The model is hallucinating frequently when retrieving internal company knowledge. Which evaluation methodology should the developer implement to quantify the ground truth accuracy of the RAG pipeline?

Accepted Answer

Implement a RAGAS-based evaluation pipeline to measure faithfulness and answer relevance.. RAGAS (Retrieval-Augmented Generation Assessment) is a framework specifically designed to evaluate RAG pipelines. It measures faithfulness (whether the answer is grounded in the retrieved context) and answer relevance (how relevant the answer is to the question). These metrics directly quantify ground truth accuracy and help identify hallucinations. Therefore, implementing a RAGAS-based evaluation is the correct approach.

Answer

Increase the temperature setting to 1.0 to encourage more creative responses.

Answer

Use Amazon CloudWatch logs to calculate the latency of every inference call.

Answer

Run a standard BLEU score evaluation against the model's training data.

Question 3

A security auditor requests a report on Bedrock usage. Which TWO of the following AWS services can be used to generate this report?

Accepted Answer

Amazon CloudWatch.. Amazon CloudWatch (C) is correct because Bedrock publishes runtime metrics and invocation logs to CloudWatch, allowing an auditor to build usage reports from model invocation counts, latency, and token metrics. AWS CloudTrail (D) is correct because it records Bedrock API activity such as InvokeModel and Converse calls as management/data events, which can be queried and exported for an audit report on who used Bedrock and when. AWS CodePipeline (A) is a CI/CD orchestration service and does not capture Bedrock usage data. AWS AppConfig (B) is for deploying and managing application configuration and feature flags, not for reporting service usage. AWS Trusted Advisor (E) provides best-practice checks and recommendations, not detailed Bedrock invocation reporting.

Answer

AWS CodePipeline.

Answer

AWS AppConfig.

Answer

AWS Trusted Advisor.

Question 4

You are deploying a foundation model on Amazon SageMaker. To optimize inference costs for an application with unpredictable traffic patterns and significant idle periods, which deployment option should you select?

Accepted Answer

SageMaker Serverless Inference. Serverless Inference is the most cost-effective solution for unpredictable traffic as it eliminates the cost of idle instances by automatically scaling resources based on request volume.

Answer

Batch Transform

Answer

Real-time inference with provisioned Auto Scaling instances

Answer

Asynchronous Inference

Question 5

You are tuning a Large Language Model on SageMaker and notice that GPU memory utilization is consistently low despite high latency. Which action is most appropriate to optimize performance?

Accepted Answer

Implementing dynamic batching in your inference handler. Dynamic batching groups multiple incoming inference requests into a single forward pass on the GPU, which raises GPU utilization and throughput while reducing per-request latency overhead. Low GPU utilization with high latency is the classic signature of under-batched inference where the accelerator sits idle between small requests. Implementing dynamic batching in the inference handler directly addresses this by keeping the GPU busy with concurrent work.

Answer

Increasing the number of instances

Answer

Switching to a larger instance type

Answer

Moving the model to CPU instances

Question 6

Which THREE actions can a developer take to mitigate hallucinations in a RAG application?

Accepted Answer

Use a clear system prompt that restricts the model to only use provided information.. Lowering temperature, providing context, and defining strict system prompts all help limit the model's creative deviation.

Answer

Use the retrieve-only approach and perform generation with a custom internal engine.

Answer

Allow the model to use its internal training data for any gap in knowledge.

Question 7

When designing a prompt for a coding assistant, which THREE pieces of information should be included to get the best results?

Accepted Answer

Frameworks or libraries to be used. Option A (frameworks or libraries to be used) is correct because specifying the exact frameworks or libraries (e.g., React, Django, Spring Boot) constrains the assistant to generate idiomatic code and APIs that match the intended stack, avoiding incompatible or generic solutions. Option B (the programming language) is correct because the language (e.g., Python, TypeScript, Go) determines syntax, standard library usage, and runtime behavior, so naming it prevents the assistant from guessing and producing code in the wrong language. Option D (specific formatting requirements) is correct because stating formatting constraints such as PEP 8, ESLint rules, indentation style, or output structure ensures the generated code conforms to project standards and is directly usable. Option C (the AWS account ID) is not relevant to prompt design for a coding assistant, since it is an environment-specific identifier that does not influence code generation. Option E (the model's internal parameter count) is not something a prompt author controls or needs to specify, as it describes the underlying model architecture rather than the task requirements.

Answer

The AWS account ID

Answer

The model's internal parameter count

Question 8

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 9

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 10

You are deploying a Generative AI application using Amazon SageMaker JumpStart. You need to ensure that the model endpoint is protected against prompt injection attacks. Which approach is most effective?

Accepted Answer

Use a prompt-based filter system that detects malicious intent via a secondary validation model. A prompt-based filter system that uses a secondary validation model to detect malicious intent is the most effective approach because prompt injection is a semantic/behavioral attack that must be detected by analyzing the input and output content, not by infrastructure controls. A dedicated classifier or LLM-as-judge can flag jailbreak patterns, instruction overrides, and data exfiltration attempts before they reach the primary model.

Answer

Encrypt the model weights with a CMK

Answer

Enable Amazon Inspector on the endpoint

Answer

Disable the API Gateway execute-api method

Question 11

Which THREE services/features are essential for establishing a secure perimeter for an Amazon Bedrock deployment?

Accepted Answer

Interface VPC Endpoints (PrivateLink). Interface VPC Endpoints (PrivateLink) (A) are essential because they let you access Amazon Bedrock privately from within your VPC over the AWS private network, keeping Bedrock traffic off the public internet and enabling granular endpoint policies. Security Groups on the VPC (D) are essential because they act as stateful virtual firewalls controlling which resources can reach the Bedrock interface endpoint on TCP 443, enforcing least-privilege network access at the perimeter. IAM Policies for resource access control (E) are essential because they authenticate and authorize which principals may invoke Bedrock models and actions, providing the identity-based control layer of the secure perimeter. AWS Public Access Block on all buckets (B) is not essential here since it governs S3 bucket exposure, not Bedrock access, and AWS Shield Standard (C) is automatically enabled DDoS protection that is not a Bedrock-specific perimeter control.

Answer

AWS Public Access Block on all buckets

Answer

AWS Shield Standard

Question 12

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 13

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

Accepted Answer

Logging and monitoring policies. Option A (Logging and monitoring policies) is correct because auditors need documented evidence of how the AI application records access, inference requests, and anomalous activity, typically via services like CloudWatch, CloudTrail, or Azure Monitor, to demonstrate continuous detection and incident response. Option B (Encryption methods for training/RAG data) is correct because it specifies the cryptographic controls protecting data at rest and in transit — for example AES-256 at rest and TLS 1.2+ in transit — which is essential for verifying confidentiality of sensitive training and retrieval-augmented generation data. Option E (Description of safety filters (Guardrails)) is correct because documenting content moderation, prompt-injection defenses, and output filtering demonstrates that the application enforces responsible-AI and abuse-prevention controls required by audit frameworks. Option C is incorrect because exposing public links to internal datasets would itself be a security and privacy violation, not an audit artifact. Option D is incorrect because storing or sharing user passwords in documentation violates credential-handling best practices; auditors review password policies and hashing (e.g., bcrypt/Argon2), never actual passwords.

Answer

Public links to all internal datasets

Answer

User passwords for all employees

Question 14

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

Accepted Answer

InvokeModel. InvokeModel is the Bedrock runtime API used to send a direct inference request to a foundation model, passing a prompt in the model-specific request body and receiving generated text. It does not require a Knowledge Base because it operates directly against the model. RetrieveAndGenerate, by contrast, is specifically designed for RAG workflows that query a Knowledge Base.

Answer

RetrieveAndGenerate

Answer

GetModelInvocation

Answer

CreateKnowledgeBase

Question 15

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

Accepted Answer

Mapping intent to an API definition. When a Bedrock Agent invokes an Action Group, it first interprets the user's natural language intent (B) to understand what the user is asking for, using the foundation model's reasoning over the conversation and the action group's description. It then maps that intent to an API definition (A), matching the identified intent to the specific operation, path, and parameters declared in the OpenAPI schema of the action group. Finally, the agent invokes the defined Lambda function (D) that backs the action group, passing the extracted parameters so the business logic can execute and return a result to the agent. Options C and E are incorrect because automatic scaling of the API endpoint is handled by the underlying compute service (e.g., Lambda or API Gateway), not by the Bedrock Agent, and the agent does not retrain the foundation model on new data as part of action group invocation.

Answer

Automatically scaling the API endpoint

Answer

Retraining the model on new data