Andrew Ng and Isa Fulford teach how to build applications with the ChatGPT API, covering chained LLM calls, external data lookups, and output evaluation. The course uses a customer service assistant as a working example and assumes Python familiarity.
I am thrilled to be back working with OpenAI's ease fold for the game to bring you a new course on building Systems with the ChatGPT API.
Thanks Andrew. Yes, I'm happy to be back and excited to get into this material with you all.
Previously we presented a course on how to prompt a language model like ChatGPT and showed you how to use the ChatGPT API in Python. But building an application requires much more than a single prompt or a single call to an LLM. In this short course, we'd like to share with you best practices for building a complex application using an LLM.
We'll use the running example of building an end to end customer service assistant system that chains multiple calls to a language model and uses different instructions depending on the output of the previous call, and sometimes even looks things up from external sources. You'll also learn how to evaluate your system's output for safety, accuracy and quality, which are all important elements of using these language models in application development. This course is accessible to anyone familiar with Python. If you have taken our previous course on ChatGPT prompt engineering for developers, it will help you dive deeper into how to use these models.
LLMs are a powerful tool for you for developers to build complex applications. We look forward to going to the details with you of how to do this and hope you enjoy the course. It.
