Build a question-answering bot using generative AI

This lab demonstrates to a how to build a question-answering chatbot that uses stateless, retrieval augmented generation to provide answers to your questions about AWS Classroom courses.

Objectives:

By the end of this lab, you should be able to do the following:

• Explain how retrieval augmented generation can be used to improve the output produced by Generative AI applications.
• Deploy a Lex chatbot powered by a large language model.
• Connect Langchain to a model launched in Amazon SageMaker.

Prerequisites:

To complete this lab it is recommended that you have a technical understanding of:

• Amazon SageMaker
• Amazon Kendra
• Amazon Lex

Being familiar with FLAN and LLMs will be a benefit

Audience:

• Solutions Architect
• Data Engineer
• Data Scientist
• Developer

Outline:

Task 1: Deploy a large language model (LLM)
Task 2: Add an Amazon Kendra data source
Task 3: Create an Amazon Lex V2 chatbot
Task 4: Query your large language model endpoint
Task 5: Implement a RAG workflow
Task 6: Deploy a web app with Cloudformation