Skip to main content

Mini-course introduction

Welcome to this step-by-step mini-course on how to build an AI agent capable of handling complex queries, like multiple color requirements.

In this course, you'll learn how to create a more advanced AI agent using Python, Pinecone, and LlamaIndex, complete with metadata filtering to make accurate recommendations.

What you'll learn​

By the end of this mini-course, you'll know how to:

  1. Vectorize product data using AWS Titan multimodal model to generate embeddings
  2. Store embeddings in a vector database by upserting data into Pinecone
  3. Build a query engine using LlamaIndex to interact with your vector database
  4. Enhance retrieval accuracy by applying metadata filtering for advanced queries
  5. Create a smart AI agent capable of handling complex, multi-attribute queries
  6. Compare naive query engines with advanced AI agents, identifying key improvements

Prerequisites​

This mini-course is designed for developers with some familiarity with Python. Here's what you'll need to follow along:

1. Python & Jupyter Notebook

  • We'll write all code in a Jupyter Notebook environment.
  • You'll use a pre-prepared dataset with shoe product data and images.

2. AWS Titan Multimodal Model

  • We'll use this model to generate embeddings (numerical representations) for both text and images.

3. Pinecone

  • A cloud-based vector database where we'll store and retrieve our embeddings.

4. LlamaIndex

  • A powerful library that allows us to query vector data and build intelligent agents.

Meet SoleMates: our example store​

This mini-course revolves around our fictional online shoe store, SoleMates.

Throughout the lessons, we'll use SoleMates as the context for all examples and tasks:

SoleMates is our fictional online shoe store

SoleMates is our fictional online shoe store

Why take this mini-course?​

  1. Practical & Hands-On: Follow along step-by-step to build a real AI agent
  2. Industry-Relevant Tools: Gain experience with AWS Titan, Pinecone, and LlamaIndex
  3. Problem-Solving Focus: Learn how to handle real-world challenges, like complex metadata filtering
  4. Clear Comparison: See how an advanced AI agent outperforms a simple query engine

Let's get started with the challenge in the next lesson.