Chatbots are useful tools for providing customers and employees access to internal documents or information. A chatbot backed by a large language model (LLM) can improve the user experience and reduce the need for human intervention. However, LLMs are often limited in their answers as they rely on pre-trained data and may not provide contextually relevant answers.

To enhance the capabilities of a chatbot and address certain limitations of LLMs, businesses can use the Retrieval-Augmented Generation (RAG) technique. RAG combines retrieval-based and generative AI models to provide more accurate and contextually relevant answers to user prompts. Embeddings represent the content of documents and user prompts as a vector store, and these embeddings are passed to an LLM to generate answers based on the additional context provided.

In this lab, you will deploy a PDF chatbot application that uses retrieval-augmented generation to answer prompts based on embeddings generated from PDF documents. The chatbot will leverage the LangChain framework and Amazon Bedrock to generate embeddings and answers. The Streamlit framework will be used to render the chatbot's user interface. You will configure and deploy the application as an Amazon ECS Service and interact with the chatbot to test its functionality.

Note: This lab utilizes a file embedding solution covered in a separate lab. It is recommended to complete the Embedding Documents With LangChain and Amazon Bedrock lab before starting this lab.

Learning objectives

Upon completion of this intermediate-level lab, you will be able to:

• Employ the Retrieval-Augmented Generation (RAG) technique to generate answers to questions based on embeddings
• Configure and deploy a PDF chatbot application to an Amazon ECS Service

Intended audience

• Candidates for the AWS Certified Machine Learning Specialty certification
• Cloud Architects
• Software Engineers

Prerequisites

Familiarity with the following will be beneficial but is not required:

• Amazon Bedrock
• AWS Fargate for Amazon ECS
• Amazon Simple Storage Service (S3)
• AWS Serverless Application Model (SAM)
• LangChain
• Streamlit

The following content can be used to fulfill the prerequisites:

• Embedding Documents With LangChain and Amazon Bedrock
• Enhancing Generative AI Models With Retrieval-Augmented Generation (RAG)
• Deploying AWS Lambda Functions With Container Images