AWS GenAI + Weaviate: Hands-on Workshop

This repository is for a hands-on workshop for building intelligent search systems, RAG workflows, and AI agents, with Weaviate vector database and AWS Bedrock in less than a day.

Caution

Optionally, you can run the workshop on your own AWS account. Doing so will incur costs on your own account. We cannot be responsible for any costs incurred on your personal AWS account. Please proceed with caution and at your own risk.

Prerequisites

None required - AWS account and development environment provided.

Recommended: Some familiarity with AWS services and Python programming.

Workshop instructions

Note

At the workshop, the instructor will guide you through the setup. If you miss any steps, you can follow the visual GuideFlow guides linked below.

AWS account & resources setup

Set up your own AWS account and resources for the workshop.

The instructor will take you through the same steps as in this GuideFlow visual guide. The steps are to:

1. Go to the provided AWS workshop link to access the temporary AWS account for the workshop
   • 2025 Sep 23
2. Download this CloudFormation template file (0-setup-weaviate.yaml).
3. Set up the AWS resources, including:
   • Access the AWS workshop account.
     • You may need to authenticate with a one-time password (OTP) sent to your email.
   • Open the AWS Management Console
   • Obtain access to the Bedrock AI models
   • Spin up a Weaviate database on AWS ECS
   • Set up SageMaker Studio where you will run the workshop notebooks

Workshop setup

1. Follow this visual guide for setting up the workshop repository. This shows you how to:
   • Set up a SageMaker Studio JupyterLab environment
   • Clone this repository into your SageMaker Studio environment

• The "Multimodal RAG" workshop is in the multimodal-rag directory, with the 0-setup.ipynb notebook.
• The "Build your own agent" workshop is in the agent directory, with the lesson-1.ipynb notebook.

Resources

• Weaviate documentation: https://docs.weaviate.io/weaviate
• Weaviate Python client: https://weaviate-python-client.readthedocs.io
• AWS Bedrock documentation: https://docs.aws.amazon.com/bedrock/latest/userguide/
• Pydantic AI documentation: https://ai.pydantic.dev/

Repository notes

• For students, most of the required packages are pre-installed in the SageMaker Studio environment.
  • The notebooks include any installation instructions for any additional required packages.
• This project was developed with uv. The primary list of required packages are in pyproject.toml; although a requirements.txt file is also provided for convenience.

Instructor / developer notes

• There are two versions of notebooks in the multimodal-rag workshop:
  • *.ipynb: The student notebooks with student TODOs
  • *-complete.ipynb: The completed notebooks with solutions
• Run generate_student_notebooks.py from the multimodal-rag directory to regenerate the student notebooks from the completed notebooks.
  • See the comments in the script for more details.