Build with AI: LLM-Powered Applications with Streamlit

Course details

In this hands-on course, instructor Megan Silvey shows you how to utilize Streamlit to build web applications. This open-source Python framework has become a key tool for data scientists and AI/ML engineers. In particular, learn how to design and deploy a retrieval‑augmented generation (RAG) document Q&A chatbot in Streamlit using Python and OpenAI’s API. Get started with an overview of Streamlit and large language models (LLMs), along with best practices for working with AI and APIs. Next, find out how to prepare text data to create a RAG pipeline that integrates into a Streamlit chat interface. By the end of this course, you’ll be prepared to test, maintain, and deploy a fully functional chatbot on Streamlit Community Cloud.

Learning objectives
Properly prepare data and build a local vector store for a Streamlit chatbot.
Implement a retrieval‑augmented generation (RAG) pipeline that retrieves context and queries an LLM for accurate answers.
Create an interactive chat UI in Streamlit, manage chat history, and handle errors.
Deploy and maintain a RAG‑powered chatbot on Streamlit Community Cloud.

Concepts

0. Introduction

01 - Build your first LLM-powered app with Python and Streamlit
02 - GitHub Codespaces

1. Hands-On - Building the Chat App Foundation

03 - Why use Streamlit to build AI-powered web apps
04 - Build your first Streamlit app
05 - Basic Streamlit commands for web development
06 - Build chat features - Add input and display messages

2. LLM Foundations

07 - What are large language models (LLMs)
08 - What is retrieval-augmented generation (RAG)
09 - Guidelines for working with AI and APIs
10 - How to connect to OpenAI API
11 - Send user prompts to an LLM and display the response
12 - Save and display chat history in your application

3. Building Your Knowledge Base

13 - How the document Q&A chatbot works
14 - Introducing Explore California
15 - Prepare text data for embedding
16 - Generate embeddings from text for searchability
17 - Create a Faiss vector store for fast retrieval
18 - Query the vector database to find relevant information
19 - Construct effective RAG prompts for better LLM answers
20 - Use the RAG query function to combine search and chat

4. Creating the Chatbot Interface

21 - Create a chat UI in Streamlit for LLM interactions
22 - Integrate the RAG pipeline into your Streamlit app
23 - Handle errors gracefully with your chatbot
24 - Provide clear and helpful feedback to users
25 - Test your chatbot to ensure it works smoothly
26 - Maintain and improve your chatbot
27 - Deploy your chatbot to Streamlit Community Cloud for free

Conclusion

28 - Next steps

Build with AI: LLM-Powered Applications with Streamlit

Authors

Megan Silvey