Welcome to DAT301! The integration of embeddings generated from LLMs for Amazon Aurora PostgreSQL-Compatible Edition and Amazon RDS for PostgreSQL presents a powerful and efficient solution for optimizing the product catalog similarity search experience. By using foundation models and vector embeddings, businesses can enhance the accuracy and speed of similarity searches by using Retrieval Augmented Generation (RAG), which ultimately leads to improved user satisfaction and a more personalized experience.
In this workshop, you'll explore building GenAI-powered e-commerce solutions using a fictitious company called Blaize Bazaar.
- Integrate foundation models with e-commerce data for product insights
- Build generative AI-powered product recommendations
- Analyze shopping trends using vector embeddings
- Create personalized customer experiences
- Streamline operations (customer support and inventory management) using Amazon Bedrock Knowledge Bases and Amazon Bedrock Agents
This workshop uses a subset of the Amazon Products Dataset (2023), which contains information about 21,704 Amazon products collected in January 2023.
Dataset Details:
- Source: Amazon Products Dataset 2023 (21,704 Products)
- Asaniczka. (2023). Amazon Products Dataset 2023 (1.4M Products) [Data set]. Kaggle. https://doi.org/10.34740/KAGGLE/DS/3798081
Before starting this workshop, you should have:
- An AWS Account with admin access
- Basic understanding of PostgreSQL and Python
Please follow the lab guide used for the workshop to get started.
This workshop environment is designed for learning and should not be used for production workloads. Consider the following security notes:
- Sample data is used and contains no sensitive information
- IAM roles are created with workshop-specific permissions
- All resources should be deleted after workshop completion
To avoid ongoing charges, delete all workshop resources.
- pgvector Documentation
- Claude API Documentation
- AWS Bedrock Documentation
- Anthropic Prompt Engineering Documentation
- Amazon Bedrock Knowledge Bases
- Bedrock Agents Guide
We welcome contributions! Please see our contribution guidelines for details.
This workshop is licensed under the MIT-0 License. See the LICENSE file.
The Amazon Products Dataset used in this workshop is subject to Kaggle's database licensing terms and Alexander Saniczka's usage terms. Please refer to the dataset page for complete licensing information.
If you discover a potential security issue, please notify AWS Security rather than opening a public issue. For other problems or suggestions:
- Open a GitHub issue
- Provide detailed description of the problem
- Include steps to reproduce if applicable
- Shayon Sanyal, Principal WW PostgreSQL Specialist SA, AWS
Copyright Amazon.com, Inc. or its affiliates. All Rights Reserved.