Data Engineering San Diego - Intro to Large Language Models
Special edition of the Data Engineering group where Data Engineering San Diego partnered with San Diego Software Engineers for an event.
We had a lot of in-person attendees and a lot of conversation, so the mic could not be situated to catch all of the audio well. We try to make the livestream helpful for those that cannot attend in person, but the room mics do not pick up enough conversation to provide as good of a virtual experience as we would like. *
Introduction to Large Language Models for Software/Data Engineers
Large Language Models (LLMs) are the foundation behind tools like GitHub Copilot and chatGPT. In this talk, we'll briefly cover the history and the technical architecture of LLMs, then jump straight into techniques on how those models improve what software products we create and how we create them.
We'll talk about Prompt Engineering, then use Copilot X to show you how to streamline developer productivity, demostrate techniques like Step-by-Step reasoning to improve LLM output, as well as explore the visible horizons of the "Capability Overhang," i.e. the LLM capabilities and features yet to be discovered. We'll discuss how these models can both transcend, disrupt, and possibly hurt both the technological and social landscapes of the future.
Will LLMs make Software Engineering obsolete as we know it? We'll explore that as well.
This is intended to be a very interactive session, please bring your ideas, thoughts and experiments to share with the group.
Don't have access to chatGPT yet? We'll make sure you get to play with it while you're here, and explore ideas together.
About the speaker
Eugene Chuvyrov is a software engineer on the Innovation team at ServiceNow and a Machine Learning/AI enthusiast.
