This course is designed for application developers interested in building generative artificial intelligence (generative AI) applications using either the Amazon Bedrock APIs or AWS-LangChain integration. In this course, you will explore the architecture patterns and implementations to support generative AI use cases such as generating and summarizing text, retrieval augmented generation (RAG), and question answering.
You learn to build RAG application using Amazon Bedrock Knowledge Bases, and AI Assistants that use knowledge bases and user-developed tools to answer questions using Amazon Bedrock Agents. You'll also learn to implement safeguards customized to your application requirements and responsible AI policies using Amazon Bedrock Guardrails.
- Course level: Advanced
- Duration: 9 hours
Activities
This course includes eLearning interactions, knowledge checks, and labs.
Course objectives
In this course, you will learn to:
- Identify the components of a generative AI application and the options to customize a foundation model (FM)
- Describe Amazon Bedrock foundation models, inference parameters, and key Amazon Bedrock APIs
- Describe the architecture patterns that can be used to build generative AI applications
- Identify Amazon Web Services (AWS) offerings that help with monitoring, securing, and governing your Amazon Bedrock applications
- Describe LangChain components such as prompt templates, chains, retrievers, and agents
- Apply LangChain components to Amazon Bedrock models to build and test use cases such as text and code generation, summarization, RAG, and question answering
- Use Amazon Bedrock Knowledge Bases to implement RAG applications using best practices
- Use Amazon Bedrock Agents with Amazon Bedrock Knowledge Bases and Amazon Bedrock Guardrails for agent applications
Intended audience
This course is intended for:
- Generative AI application developers
Prerequisites
We recommend that attendees of this course have:
- Intermediate to expert-level proficiency with Python programming language
- AWS Technical Essentials (Fundamental)
- AWS Lambda Foundations (Fundamental)
- Amazon Bedrock Getting Started (Fundamental)
- Foundations of Prompt Engineering (Intermediate)
Course outline
Module 1: Introduction to Amazon Bedrock
- Building Generative AI Applications on Amazon Bedrock
- Applications and Use Cases
- Topics Covered in Future Modules
- Conclusion
Module 2: Application Components
- Overview of Generative AI Application Components
- Foundation Models and the FM Interface
- Working with Datasets and Embeddings
- Additional Application Components
- RAG
- Model Fine-Tuning
- Securing Generative AI Applications
- Introduction to Architecture Patterns
- Test Generation and Text Summarization
- Question Answering
- Chatbots
- Code Generation
- LangChain and Amazon Bedrock Agents
- Knowledge Check
- Conclusion
Module 3: Foundation Models
- Introduction to Amazon Bedrock Foundation Models
- Using Amazon Bedrock FMs for Inference
- Amazon Bedrock Methods
- Data Protection and Auditability
- Knowledge Check
- Conclusion
Module 4: Using LangChain
- Optimizing LLM Performance
- Integrating AWS and LangChain
- Using Models with LangChain
- Constructing Prompts
- Structuring Documents with Indexes
- Storing and Retrieving Data with Memory
- Using Chains to Sequence Components
- Managing External Resources with LangChain Agents
- Knowledge Check
- Conclusion
Lab 1: Explore Generative AI Use Cases using LangChain and Amazon Bedrock
- Introduction
- Task 1a: Perform Text Generation
- Task 1b: Perform Text Generation using a prompt that includes Context
- Task 2a: Text summarization with small files with Titan Text Premier
- Task 2b: Abstractive Text Summarization
- Task 3: Use Amazon Bedrock for Question Answering
- Task 4: Conversational Interface - Chat with Llama 3 and Titan Premier LLMs
- Task 5: Invoke Bedrock model for code generation
- Task 6: Bedrock model integration with LangChain Agents
- Conclusion
Module 5: Using Knowledge Bases
- Overview of RAG
- RAG Use Cases
- RAG Architecture
- Challenges While Building RAG Applications
- Responsible AI
- Amazon Bedrock Knowledge Bases
- Data Ingestion Into Knowledge Bases
- Fully Managed RAG using Amazon Bedrock Knowledge Bases
- Customized RAG Using Retrieve
- Evaluating RAG Applications
- Best Practices and Advanced Techniques for RAG
- Amazon Bedrock Guardrails
- Knowledge Check
- Conclusion
Lab 2: Build and Evaluate Retrieval Augmented Generation (RAG) Applications Using Amazon Bedrock Knowledge Bases
- Introduction
- Task 1: Leverage a fully-managed RAG application with Amazon Bedrock's RetrieveAndGenerate API
- Task 2: Build a Q&A application using Amazon Bedrock Knowledge Bases with Retrieve API
- Task 3: Test the Query Reformulation process supported by Amazon Bedrock Knowledge Bases
- Task 4: Build and evaluate Q&A Application using Knowledge Bases using RAGAS framework
- Task 5: Test the guardrail functionality on Knowledge Bases using RetrieveAndGenerate API
- Conclusion
Module 6: Using Agents
- Introduction to Agents
- Use Cases for Agents
- Overview of Amazon Bedrock Agents
- Creating and Deploying Agents
- Creating and Deploying Agents in the AWS Console
- Creating and Deploying Agents Programmatically
- Creating Agent Action Groups
- Action Group Invocation Types
- Amazon Bedrock Agents Integrations
- Amazon Bedrock Knowledge Bases
- Amazon Bedrock Guardrails
- Deploying and Invoking Agents
- Knowledge Check
- Conclusion
Lab 3: Explore Amazon Bedrock Agents integrated with Amazon Bedrock Knowledge Bases and Amazon Bedrock Guardrails
- Task 1: Configure the Agent's Short-Term Memory
- Task 2: Setup the SageMaker Studio Environment
- Task 3: Test and Trace the Agent Steps With and Without Amazon Guardrails
- Conclusion
Keywords
- Gen AI
- Generative AI
