AIP-C01 Practice Exam 4 — 20 Questions

Question 1

Which THREE metrics are best used for setting up auto-scaling policies on a SageMaker endpoint?

Accepted Answer

InvocationsPerInstance. InvocationsPerInstance, CPUUtilization, and GPUUtilization are the standard metrics to trigger scaling actions.

Answer

IAMUserLoginCount

Answer

S3RequestLatency

Question 2

Which THREE of the following are components of a robust AI governance framework in an AWS environment?

Accepted Answer

Implement Amazon Bedrock Guardrails. Option A is correct because Amazon Bedrock Guardrails provides configurable safeguards such as content filters, denied topics, and PII redaction that enforce responsible-AI policies at inference time, which is a core control of an AI governance framework. Option D is correct because Bedrock Model Invocation Logging captures invocation metadata (and optionally full request/response payloads) to CloudWatch Logs or Amazon S3, delivering the auditability and traceability that governance requires. Option E is correct because IAM roles with least-privilege access restrict who and what can invoke models or manage AI resources, satisfying the access-control and segregation-of-duties principles of governance. Option B does not belong because granting all users AdministratorAccess violates least privilege and destroys accountability. Option C does not belong because publicly hosting a model endpoint exposes it to unauthorized use and data exfiltration, directly undermining governance and security.

Answer

Provide all users with AdministratorAccess

Answer

Publicly host the model endpoint

Question 3

A developer wants to ensure that the RAG pipeline is resilient to outages. What architectural pattern should be used?

Accepted Answer

Multi-region deployment with cross-region replication for the vector database.. A multi-region deployment with cross-region replication of the vector database ensures that if one AWS region becomes unavailable, the RAG pipeline can fail over to a replica region that already holds the embedded document vectors, preserving retrieval and generation. This directly addresses resilience to outages by removing the single-region failure domain for both compute and the vector store. Cross-region replication keeps the vector index synchronized so failover does not require re-ingesting and re-embedding the corpus.

Answer

Single-node deployment.

Answer

Hardcoding API endpoints.

Answer

Using only one S3 bucket.

Question 4

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

Accepted Answer

Auto Scaling policies based on custom metrics. Instance type, auto-scaling configuration, and model optimization (like quantization) are critical levers for cost efficiency.

Answer

CloudTrail log retention period

Answer

Number of S3 buckets used

Question 5

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

Accepted Answer

Apply metadata filters to control document access. Option A is correct because applying metadata filters at query time restricts retrieval to only the documents a given user is authorized to see, enforcing document-level access control within the RAG pipeline. Option B is correct because IAM-based connectors authenticate to data sources using identity and access management policies, ensuring that only permitted documents are ingested and that access respects existing enterprise permissions. Option C is wrong because a single shared user removes per-user accountability and grants everyone the same access, defeating least privilege. Option D is wrong because making all documents globally readable exposes sensitive content to every user and query. Option E is wrong because disabling indexing of sensitive documents doesn't secure them within RAG; it merely removes them from retrieval, which is a data-availability decision rather than an access-control mechanism.

Answer

Use a single shared user for all data source connections

Answer

Make all documents globally readable

Answer

Disable indexing of sensitive documents

Question 6

Your organization needs to ensure that only authorized IAM users can access specific foundation models in Bedrock. Which configuration is required?

Accepted Answer

Attach an IAM policy with bedrock:InvokeModel allowing access to specific model ARNs. Amazon Bedrock integrates with AWS IAM to control access to foundation models. To restrict which IAM users can invoke specific models, you attach an IAM policy that includes the bedrock:InvokeModel action and specifies the ARNs of the allowed foundation models. This policy can be attached to users, groups, or roles, and it enforces least privilege at the model level. Without such a policy, users either have no access or overly broad access, so this is the required configuration.

Answer

Enable AWS Organizations SCPs only

Answer

Use Amazon Cognito user pools

Answer

Use Resource-Based Policies on the foundation model

Question 7

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

Accepted Answer

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

Answer

Model evaluation dataset.

Answer

External firewall rules.

Question 8

A company is building a RAG application on Amazon Bedrock. They need to store embeddings for a large-scale knowledge base with low-latency similarity search and full integration with LangChain. Which vector database should the developer recommend?

Accepted Answer

Amazon OpenSearch Serverless vector engine. Amazon OpenSearch Serverless vector engine is purpose-built for storing and querying high-dimensional embeddings with k-NN similarity search, offering low-latency retrieval at scale and native LangChain integration via the OpenSearch vector store class. It removes cluster management overhead while supporting the vector workloads RAG applications require.

Answer

Amazon DynamoDB with global secondary indexes

Answer

Amazon Aurora PostgreSQL without pgvector

Answer

Amazon S3 with Athena query engine

Question 9

Which TWO of the following strategies are best for reducing the cost of SageMaker training jobs?

Accepted Answer

Using Managed Spot Training. Using Managed Spot Training and instance-appropriate types are standard methods for reducing training costs.

Answer

Keeping training clusters running

Answer

Increasing the number of instances

Answer

Enabling detailed logging for every epoch

Question 10

A company is using RAG with Amazon Bedrock Knowledge Bases. They need to ensure that users can only retrieve documents they are authorized to see. How should this be implemented?

Accepted Answer

Use the filters attribute in the Retrieve API call based on document metadata. Amazon Bedrock Knowledge Bases support metadata filtering on the Retrieve and RetrieveAndGenerate APIs. By attaching authorization metadata (e.g., department, clearance level) to documents during ingestion and passing a filters attribute in the Retrieve call based on the user's identity, the application ensures users only retrieve documents they are authorized to see. This enforces document-level access control at query time.

Answer

Encrypt the documents with different KMS keys

Answer

Create separate Knowledge Bases for every user

Answer

Apply IAM policies to the S3 bucket

Question 11

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

Accepted Answer

Configuring encryption for the data at rest.. Data classification, encryption configuration, and IAM management are the customer's responsibility.

Answer

Maintaining the physical security of the data center.

Answer

Patching the underlying foundation model code.

Question 12

What is the primary benefit of deploying a JumpStart model on a Multi-Model Endpoint (MME)?

Accepted Answer

It optimizes resource utilization by hosting multiple models. A Multi-Model Endpoint hosts multiple models behind a single endpoint and dynamically loads/unloads them into shared container memory on demand. The primary benefit is resource utilization: you avoid provisioning a separate endpoint per model, so many low-traffic models share one instance's CPU/GPU and memory. This is the defining advantage of MME over single-model endpoints.

Answer

It improves inference latency for a single model

Answer

It simplifies the model training process

Answer

It provides higher availability than single-model endpoints

Question 13

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?

Accepted Answer

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

Answer

Amazon CloudWatch Metrics for Bedrock.

Answer

AWS Glue DataBrew.

Answer

Amazon SageMaker Model Monitor.

Question 14

A company is building a RAG application using Amazon Bedrock and needs to ensure that sensitive PII is filtered out before it reaches the foundation model. Which feature should be used to achieve this?

Accepted Answer

Amazon Bedrock Guardrails. Amazon Bedrock Guardrails allows users to configure content filters, including PII redaction, to prevent sensitive information from being processed by or returned from foundation models.

Answer

AWS WAF on the Bedrock API endpoint

Answer

Amazon Macie on the S3 bucket housing the source data

Answer

AWS Glue DataBrew

Question 15

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?

Accepted Answer

Check CloudWatch Logs for the associated Action Group Lambda functions.. Option C is correct because Bedrock Agents invoke Action Group Lambda functions to fulfill API operations, and their execution logs, errors, and timeouts are captured in CloudWatch Logs, making them the primary place to isolate backend failures. Option D is correct because the Agent's Trace output shows the step-by-step orchestration (pre-processing, orchestration, knowledge base/action group invocation, and post-processing), revealing exactly which stage or invocation failed. Option A is wrong because Bedrock is a fully managed service with no customer-accessible restart operation, and restarting would not diagnose the root cause. Option B is wrong because raising temperature to 1.0 only increases response randomness and would not isolate an error. Option E is wrong because base foundation models cannot be retrained by customers; customization is done via fine-tuning or continued pre-training, which is unrelated to troubleshooting.

Answer

Restart the Bedrock service.

Answer

Increase the model's temperature to 1.0.

Answer

Re-train the underlying base model.

Question 16

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?

Accepted Answer

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

Answer

AWS CloudTrail logs

Answer

AWS WAF on the API Gateway

Answer

IAM Policy evaluation for Bedrock

Question 17

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?

Accepted Answer

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

Answer

Model Training Loss

Answer

Inference Latency

Question 18

When using a Knowledge Base, what is the 'Base Prompt' used for?

Accepted Answer

To instruct the model on how to format the answer using retrieved context.. The base prompt serves as the template for the LLM to understand how to use retrieved context to answer the user's question.

Answer

To define the database schema.

Answer

To authenticate the user.

Answer

To store the conversation history.

Question 19

A company is concerned about model provider policy compliance. Where can they find the specific data processing terms for their chosen model?

Accepted Answer

In the AWS Service Terms and individual model provider documentation.. The AWS Service Terms and the Bedrock documentation page specify the data handling and privacy commitments for each model provider.

Answer

In the Amazon Bedrock console 'Model Evaluation' tab.

Answer

By asking the model during an inference session.

Answer

In the Amazon CloudWatch monitoring dashboard.

Question 20

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

Accepted Answer

Amazon Bedrock Knowledge Bases. Amazon Bedrock Knowledge Bases implements Retrieval Augmented Generation (RAG) by connecting foundation models to trusted data sources stored in Amazon S3 or other supported repositories. It chunks, embeds, and indexes documents into a vector store, then retrieves relevant passages at inference time so the model's response is grounded in actual company content rather than parametric memory. This directly reduces hallucinations by supplying factual context in the prompt.

Answer

Amazon SageMaker Feature Store

Answer

AWS Glue Data Catalog

Answer

Amazon DynamoDB