Level up LLM applications development with LangChain and OpenAI

Course details

Dive into the world of large language models (LLMs) with a focus on integrating them into practical applications utilizing OpenAI APIs. Discover how to enhance LLMs with retrieval components, deploy interactive chat applications, and construct multi-retriever agents for advanced data handling. Join instructor Sandy Ludosky to gain the skills to create intelligent agents capable of performing complex tasks, from semantic searches to question-answering chatbots, significantly enhancing user experiences. Whether you're aiming to innovate in your current role or embark on new AI projects, this course provides the foundational knowledge and practical skills needed to harness the power of LLMs effectively.

Learning objectives
Learn how to interface easily and efficiently with large language models (LLM).
Learn how to add a retrieval-augmented generation (RAG) module.
Deploy LangChain chains as a REST API with LangServe.
Deploy to the cloud with Replit and Streamlit.

Concepts

0. Introduction

01 - Level up LLM applications
02 - What you should know

1. LangChain Basics - Intro to Building LLM-Powered Apps

03 - Setup and installation
04 - Create a chain and interface with LLM
05 - Define and structure a prompt
06 - Create and invoke a chain (LCEL syntax)
07 - Work with output parsers

2. Adding Similarity Search and Context

08 - Quickstart - Installation and setup
09 - Create embeddings from text (Faiss)
10 - Querying the vector store
11 - Querying as a retriever

3. Leveraging LLMs with LangChain and RAG

12 - RAG - Overview and architecture
13 - Breaking down the RAG pipeline
14 - Project setup
15 - Load and split documents into chunks
16 - Initialize a vector store (Chroma) and ingest documents
17 - Create the chain - Prompt + model + parser
18 - Create the chain - Add context with a retriever
19 - Pass data with RunnablePassthrough and query data
20 - Challenge - Create a custom agent with history
21 - Solution - Add a chain with chat history
22 - Solution - Context- and history-aware chatbot

4. Create an Interactive Web App (Streamlit)

23 - Set up the Streamlit application
24 - Build the layout with Streamlit components
25 - Adding functionality with Streamlit
26 - Challenge - Deploy your Streamlit app
27 - Solution - Add app to GitHub
28 - Solution - Deploy your app

5. Build a Q&A Agent with Multiple Data Sources and Query Analysis

29 - Retrieval with query analysis
30 - Connect to a data source and create an index
31 - Set up query analysis to handle multiple data sources
32 - Retrieval with query analysis
33 - Challenge - Retrieval with multiple data sources
34 - Solution - Q&A with multiple data sources

6. Perform Semantic Search Using MongoDB Atlas Vector Search and OpenAI

35 - Getting started with MongoDB - Create an account
36 - Build and deploy a free cluster
37 - Set up the MongoDB environment and connect to the cluster
38 - Create a secured database access (user)
39 - Load sample data and create the vector store
40 - Create the Atlas Vector Search index
41 - Run vector search queries

7. Interact with a NoSQL Database (MongoDB)

42 - Create a retrieval chain - Define the prompt
43 - Create a retrieval chain - Define the context
44 - Create a retrieval chain - Parse and format results
45 - Query documents and generate extended responses

8. LLM Fine-Tuning with the OpenAI Tools and Functions

46 - Using agents to perform actions in chains
47 - Define tools
48 - Select the perfect prompt
49 - Bind tools and create agent
50 - Create and run the agent executor
51 - Challenge - Create a multitask agent
52 - Solution - Define tools and functions

9. Deploy Chains as a RESTful API with LangServe

53 - Introducing LangServe - Installation and setup
54 - Create a server
55 - Create the routes and the endpoints
56 - Create a runnable to combine a prompt, a model, and output
57 - Challenge - Deploy a RESTful API
58 - Solution - Deploy a RESTful API

10. Finish Line - Deploy to the Cloud and Share with the World

59 - Manage and deploy an app on Render
60 - Create a GitHub repository and push your project
61 - Deploy a new web service on Render

Conclusion

62 - Conclusion