AIP-C01 Study Guide

Amazon Bedrock Knowledge Bases is designed for building and deploying knowledge bases as an additional abstraction to simplify the RAG ingestion and retrieval processes, enabling faster development cycles.

During knowledge base evaluation jobs, Amazon Bedrock makes a temporary copy of your data and deletes it after the job finishes, ensuring data privacy and minimizing storage costs.

Customize models in Amazon Bedrock with your own data using fine-tuning and continued pre-training. Alignment methods for LLMs ensure models behave according to organizational policies and ethical guidelines.

To benefit from optimizations, host your LoRA customized models on Amazon SageMaker AI or Amazon Bedrock. Implementing multi-LoRA techniques enables efficient serving of dozens of fine-tuned models.

Amazon Bedrock now extends reinforcement fine-tuning (RFT) support to popular open-weight models, including OpenAI GPT-OSS and Qwen models, expanding customization capabilities.

You can set up an S3 destination for logging in Amazon Bedrock by creating an S3 bucket where the logs will be delivered and adding a bucket policy to it.

The AWS SDK for SAP ABAP adds the audioDataDeliveryEnabled boolean field to the Model Invocation Logging Configuration, providing granular control over audio data delivery in logging.

This document details adding a guardrail component to filter content in an Amazon Bedrock flow app, including enabling content filters, applying denied topic filters, and configuring word filters.

To set up a role with permissions for guardrails, create an IAM role and attach the required permissions by following the steps at Creating a role, ensuring proper access control.

For example, you can use Amazon Bedrock Guardrails to detect and filter harmful content. The recommended guardrails include Input filters, Content filters, and Output filters.

Amazon Bedrock Guardrails enables you to implement safeguards in generative AI applications customized to your specific use cases and responsible AI policies using IAM policy-based controls.

The Amazon Bedrock Guardrails feature constrains agent behavior and enforces enterprise policy. Capabilities of guardrails include Content filtering, Topic filtering, and Word filtering.

Content can be filtered with AWS services like Amazon Bedrock Guardrails or custom validation logic. Apply these filters as a preprocessing step in your workflow, and pass filtered data to downstream components.

The audio summarization solution uses Whisper Large V3 Turbo from Amazon Bedrock Marketplace for transcription, Anthropic Claude Sonnet 3.5 for summarization, and Amazon Bedrock Guardrails for PII redaction.

The audio summarization workflow implements Amazon Bedrock Guardrails using the ApplyGuardrail API with 'OUTPUT' parameter for redacting PII in summaries, replacing sensitive data with type indicators like {PHONE} or {EMAIL}.

Facing issues in deploying an endpoint on SageMaker later to be invoked by a downstream service - ECS. Check container configuration, model artifact integrity, and IAM role permissions.

The actual problem is that the endpoint is not able to ping the container. It is because, when there are multiple containers, every container is using some ports, causing port conflicts and connectivity issues.

Logs not coming for endpoint being created in SageMaker. The status of the endpoint goes to failed with the message 'The primary container for production variant...' indicating container startup failures.

Troubleshoot missing CloudWatch metrics in SageMaker AI by checking the correct CloudWatch namespace: The AWS/SageMaker namespace includes model loading metrics from API calls to InvokeEndpoint.

Log group not found: Disable, and then re-enable your monitor. Enabling a monitor restarts log group creation, which might correct the problem. PutLogEvents permission issues can also cause logging failures.

Use the following information to help you troubleshoot errors that might come up when you're using Amazon EMR. Connection failure – If the connection fails with timeout or authentication errors, check network connectivity and IAM permissions.

To resolve this issue, use ModelStep to register the model. You can also use Boto3 or SageMaker AI Studio to register your model, ensuring proper metadata and artifact linking.

If you run into an error while creating a model evaluation job, use the following list to troubleshoot your evaluation. Check IAM permissions, KMS key access, and S3 bucket policies.

This error occurs when SageMaker updates all models to use uncompressed model artifacts. To resolve this issue, set your notebook model_version attribute to an older compatible version.

This section summarizes errors that you might encounter and what to check if you do. Permissions issues If you encounter an issue with permissions to access an AWS resource, verify IAM role trust policies and attached permissions.

Every AWS KMS key must have exactly one key policy. The statements in the key policy determine who has permission to use the AWS KMS key and how they can use it, ensuring proper encryption control.