Andrew Ng and Adrian Gonzalez Sanchez teach you to build an LLM agent that queries SQL databases and returns answers in plain English. You'll use LangChain, Azure OpenAI, and the Assistants API to turn natural language into database queries.
Hi, I'm delighted to introduce Building your own database agent created in partnership with Microsoft and taught by Adrian Gonzalez Sanchez. An LLM based agent can autonomously decide on and execute a sequence of actions. In this course you go through the process of building an agent that can interact with and extract information from tabular data and SQL databases to answer questions. By the end of this course you have built an AI agent that can search through structured data like tablet data or SQL databases. Say you have a dataset of COVID cases and you want to find out how many Covid cases there were in some state, but you want to filter by the severity of the case. You learn in this course how to use the LangChain Agents framework to map from an English query to one or more database access queries, to retrieve the information you want, and then finally to generate a natural language answer to your question. As part of this, we'll also deep dive into the Azure OpenAI function calling feature, which lets GPT4 decide when to call a function. You also learn to use the Azure OpenAI assistance feature, which is stateful, and the Code interpreter which lets you execute generated code on I'm delighted to introduce instructor Adrian Gonzalez Sanchez. Adrian is a data and AI specialist at Microsoft and he's working on cloud architectures for generative AI at Microsoft and is also author of the O'Reilly Azure OpenAI book.
Thanks Andrew. In this course you'll learn how to build your own AI agents using the Azure OpenAI service. Agents are one of the most exciting developments in LLMs right now, and I look forward to sharing with you the key building blocks like the function calling and more importantly, how to put them together. As Andrew mentioned, we'll be illustrating this concept with a database agent, but you will be able to use what you learn to build other types of agents as well. We'll start by setting up a RAG pipeline and connect it to first a CSV dataset and then a SQL database. Then you will also learn to build a function calling system and integrate the Azure OpenAI Assistance API with your SQL database agent.
With this course, I hope that whatever you're working on, you'll be better able to serve as an agent of change. Let's get started.
Good one Andrew and everyone, I hope you enjoy this class.
