AWS Bedrock Workshop (Build a Gen AI Chatbot) - Level 100

Discover Amazon Bedrock, a fully managed, serverless API for foundational models from leading AI startups. Build and customize chatbots and generative AI apps without managing infrastructure.

Discover the updated Amazon Bedrock console, automatic model access, and the model catalog. Compare Titan Text Express and other models in the chat playground, and explore agents and token counting.

Explore AWS Bedrock pricing, compare on-demand and batch models, and understand token-based costs per thousand input and output tokens; note region availability and Titan models pricing.

Explore prerequisites for building a gen AI chatbot on AWS Bedrock, including setting up an AWS account, Visual Studio Code, Python, Streamlit, Boto3, LangChain, and AWS profile authentication.

Set up a free AWS account and explore the free tier, including EC2, S3, and RDS, by creating the root user and verifying your credit card for identity.

Set up Visual Studio Code on Windows 11, install Python 3.12.2 with environment variable, verify integration in the VS Code terminal, and prep for installing boto3, Lang chain, and Streamlit.

Build a web chatbot using Streamlit for the front end, integrate Lang Chain with Boto3 and AWS Bedrock, and manage memory with a conversation chain and buffer memory.

Install boto3, streamlit, and langchain with pip in Visual Studio Code, verify installations, and launch a local streamlit app to prototype your chatbot.

Install the AWS CLI from binaries, verify the version, and enable VS Code extensions AWS Toolkit, Boto3, and Codewhisperer for smooth development and upcoming authentication setup with an IAM role.

Configure an AWS profile in Visual Studio Code using either an IAM user with access key (not recommended) or the IAM Identity Center, generate and securely manage Bedrock access keys.

Set up the AWS CLI in VSCode by configuring access key, secret and region for the bedrock profile, verify with a test S3 access, and prepare for bedrock code generation.

Use Identity Center SSO with an SSL profile to authenticate to AWS on every session, create Identity Center users, attach permission sets, and access Bedrock resources in the AWS console.

Configure the AWS CLI to use SSO via IAM Identity Center and authorize the device. Switch profiles to grant S3 read access and validate by listing an S3 bucket.

Explore chatbot architecture, including context from uploaded data via a database and retrieval augmented generation for accurate responses, then memory from chat history with LangChain and Titan LM.

Develop a Python-based chatbot backend (part 1) using bedrock chat with lang chain, backed by a conversation summary buffer memory, and build a streamlit front end to deploy a chatbot.

Define a conversation summary buffer memory with the Titan LM to summarize past messages via token-based flushing, then connect it to a conversation chain and a Streamlit front end.

Create a frontend chat interface with Streamlit and a backend glib, initialize memory and chat history, render user and bot messages, and handle chat input with get chat response.

Learn how to run a Streamlit chatbot from the terminal, fix memory and transformers errors, and explore the conversation summary buffer memory with LangChain to maintain context.

Gain insight into chat bot architecture and key components, and explore text-to-text bots alongside image, music, and story generation. Learn to build advanced, fine-tuned bots using your own data.