LLMOps Applications

In this course, you will learn how to interact and create applications with Generative AI APIs like OpenAI. We will cover how to set up GitHub Codespaces with GitHub Copilot, learn about DALL-E and use it to generate images from text prompts, learn how to fine tune LLM models to a specific use case and check for accuracy to avoid hallucinations. You will learn the capabilities of the Whisper API for encoding and summarizing audio and video files. We will walk through an example of ingesting a speech audio file, getting the transcript from OpenAI, and then getting a summary of the key points back. This provides the basis for applications like meeting or video summarizers. Additionally we will cover the rapidly growing GenAI market, key partnerships and acquisitions happening in the space, as well as important principles around AI ethics and safety when developing AI applications.

Learning Objectives

Set up and use GitHub Codespaces and Copilot to optimize programming Understand and use the DALL-E API to generate images from text Fine tune OpenAI models to a specific use case (ex. Reddit Q&A bot) Use the Whisper API to transcribe and summarize audio and video files Get an understanding of the fast-growing GenAI market and key players Learn AI ethics principles for developing responsible AI applications

Course Content

Setting up your development environment with GitHub Codespaces and Copilot Using the DALL-E API to generate images from recipe text and shopping lists Data preparation and fine-tuning an OpenAI model on Reddit Q&A data Using the fine-tuned model as a Reddit Q&A assistant Transcribing and summarizing audio/video with the Whisper API Overview of the GenAI market size, growth projections, and key partnerships Important AI ethics principles to consider when developing AI applications

About your instructor

Derek Wales is a West Point Graduate (Electrical Engineering/Track) who served ten years in the Army as an Engineer Officer before being medically retired. While on active duty, he designed a laser target location module with the Johns Hopkins Applied Physics Laboratory, served as a NATA Company Trainer in Germany, and Commanded a Basic Training Company. Since leaving the military, Derek has graduated from the Duke Data Science (MIDS) and the Duke Weekend Executive MBA (WEMBA) Programs. He worked as a data scientist and transitioned into product management at a large tech company, while teaching at the MIDS Program every semester (covering various topics from virtualization to GenAI) and the Business School (running Data Center Tours).

Resources

• Hugging Face for MLOps

• Doing MLOps with Databricks and MLFlow

• Practical MLOps book