Build Real world End-to-End AI Agents using AWS Bedrock

Discover how retrieval augmented generation links LLMs with external knowledge in real time, using vector databases and embeddings to improve factual accuracy and enable trustworthy AI.

Understand function calling and orchestration to let a language model call external tools via Python functions, coordinating data from MySQL, MongoDB, and a vector store on AWS Bedrock.

Discover agentic AI; autonomous, goal-driven agents that reason, adapt, and act on behalf of users to automate tasks across data sources, function calls, and orchestration, with AWS Bedrock.

Install essential local tools to follow labs: Visual Studio (or any editor), Python, Jupyter Notebook, Streamlit, Boto3, Docker, and the AWS CLI; deploy Streamlit chatbots to AWS ECS.

Install and verify AWS CLI and Docker Engine across Linux, macOS, and Windows; configure AWS with access and secret keys, and ensure the Docker daemon runs in the background.

Discover how AWS Bedrock provides a fully managed serverless platform for building generative AI apps, with unified access to foundation models, fine-tuning, RAG, and AI agents via a single API.

Explore the AWS Bedrock console, navigate the model catalog and marketplace deployments, and manage builders, prompts, and access regions to deploy end-to-end AI agents.

Get started with Python and AWS Bedrock using Boto3, running code in a local Jupyter notebook and invoking a serverless LLM via a unified API.

Deploy a streamlit chatbot as a dockerized app to aws ecs using ecr, containerize and push the image, and run bedrock-backed model invocations.

Learn to invoke a multimodal Bedrock stability diffusion model via a unified API to generate images from text prompts, and store results in S3 with a pre signed URL.

Convert IPython notebook image generator diffusion code into a streamlit chatbot, test locally, deploy a dockerized chatbot to aws ecs with s3 images accessible via pre signed URLs using ChatGPT.

Discover how AWS Bedrock knowledge bases bind private data to LLMs via a managed vector store with session management. The course uses OpenSearch Serverless and explains token and compute pricing.

Learn to create a knowledge base with a vector store in AWS Bedrock by uploading PDFs to S3, configuring an IAM user, and selecting Titan embeddings v2 with OpenSearch Serverless.

Sync the knowledge base with data sources and test it with titan text g1 on demand; validate vector store responses and deploy rack pipeline with bedrock python sdk to aws.

Invoke a knowledge base with Python SDK and AWS Bedrock, configure models and Lambda API gateway, and query a vector store (OpenSearch) for precise answers.

Build and deploy an intelligent chatbot grounded to a vector store for EV charging infrastructure, using PDF content and AWS Bedrock LMS, with scalable deployment and API or UI access.

Set up a knowledge base with a vector store in bedrock and an s3 data source for EV charging. Deploy a lambda function exposed by api gateway to enable retrieval.

Deploy a streamlit chatbot to AWS ECS using a Docker image, pushing to ECR, and running a Fargate task, enabling an LLM-driven conversational interface.

Build Bedrock agents to automate complex tasks by interacting with various systems and services on AWS, using foundation models via Bedrock APIs and SDKs to run autonomous, end-to-end workflows.

Build a Bedrock powered product inventory assistant that queries DynamoDB and OpenSearch knowledge base data, with Bedrock handling memory and orchestration, deployed serverless via a Streamlit frontend.

Set up a DynamoDB table product_inventory and ingest inventory items CSV, then create a Bedrock knowledge base with an S3 data source and a vector store using item text embeddings.

Deploy and test a Python lambda that queries DynamoDB with department, brand, and category filters, returning matched products, and prepare for Bedrock integration.

Create a Bedrock agent, attach a knowledge base and a Lambda action group, define an API schema, and test multi-source queries against DynamoDB and the knowledge base.

Invoke the agent with the Python Boto3 SDK by creating a session and alias, then call invoke_agent to fetch results from DynamoDB. Clean up by deleting knowledge bases and agents.

Learn to deploy multiple subagents and a supervisor in Bedrock to orchestrate parallel tasks, route queries, and consolidate responses for real world ai workflows.

Develop a mortgage assistant using generative AI that handles new applications and policy queries, retrieves existing loan details, with agents using DynamoDB tables and knowledge bases, overseen by a supervisor.

Set up the first sub-agent for mortgage applications by deploying a lambda function and creating a DynamoDB table mortgage_applications, integrated with AWS Bedrock.

Integrate the deployed lambda with the agent using an action group and OpenAPI schema, upload the YAML to S3, and configure a knowledge base in Bedrock for mortgage processing.

Create and ingest a DynamoDB existing mortgages table from a csv, deploy a customer-id driven lambda, and integrate it with Bedrock, a knowledge base, and testing workflow.

Set up the supervisor agent to coordinate mortgage workflows with multi-agent collaboration, using the cloud 3.5 sonnet v2 model; configure two subagents with aliases and supervisor mode routing.

Deploy a mortgage assistant chatbot to AWS ECS with Streamlit UI, dockerized deployment to ECR, alias creation, and supervisor agent integration.

Explore how to build a hotel booking assistant with AWS Bedrock, using three Lambda functions and DynamoDB to check real-time availability, query bookings, and create reservations via natural language input.

Deploy three Python Lambda functions on AWS Bedrock to check hotel bookings, verify availability, and initiate bookings, using DynamoDB booking details and room inventory tables.

Configure action groups for three lambda functions using open API schemas, upload to S3, and deploy an AI agent to check bookings, room availability, and bookings in DynamoDB.

Leverage structured data as a knowledge base with Redshift serverless, integrating with AWS Bedrock, S3, Glue, Athena, and QuickSight for analytics and machine learning data via SageMaker.

Create a Redshift serverless namespace and workgroup, configure an IAM role with S3 access, ingest rental listings from a CSV into a table, then set up Bedrock vector store.

Set up a Redshift knowledge base in AWS Bedrock, configure an IAM role, connect to Redshift serverless, sync data, grant select on rental listings, then create and attach an agent.

Configure a rental assistant agent in Amazon Bedrock, using lm cloud 3.5 v2, with a Redshift rental listings knowledge base and prepared prompts to enable qa via Python sdk.

Avoid using Redshift as a Bedrock knowledge base vector store; it can trigger unbounded SQL queries from natural language inputs, raising costs by restricting tables and databases.

Create a bedrock agent by deploying a lambda function that queries the Redshift rental listings table, configure secrets, and publish an OpenAPI YAML for bedrock integration.

Learn to build generative AI applications with Bedrock Flows, a visual builder tool that accelerates creation, testing, and deployment of custom workflows through the web console.

Build your first end-to-end workflow in Amazon Bedrock flows for ecommerce user feedback, adding categorization, pricing feedback, and feature request prompts with conditional routing.

Build real-world AI agents by integrating Bedrock knowledge bases with flows to deliver reliable responses, leveraging pricing justification and feature roadmap knowledge bases.

Explore multi-turn conversation in AWS Bedrock by using agents, enabling memory, and routing user queries to knowledge bases, with a hands-on setup using a hotel booking assistant.

Deploy agents with bedrock flows to route hotel booking inquiries and local area recommendations via a prompt-based classifier, attach a knowledge base, and enable memory for multi-turn conversations.

Deploy end-to-end ai agents with Bedrock flows to categorize user queries into hotel bookings or area recommendations, using a hotel booking agent and a knowledge base synced from S3.

Learn to invoke Bedrock flows using the Python/Boto3 SDK, manage single and multi-turn conversations with flow and alias IDs, and handle streaming responses with execution IDs.