Session: How we built and use Cody, an open source AI designed for large, messy codebases
LLMs can autocomplete functions and generate simple apps from scratch, but how much benefit do they bring in the large, messy codebases that most devs work in day-to-day?
Beyang Liu, CTO of Sourcegraph, will talk about some of the challenges his team encountered in making AI work well with real-world codebases, and how they addressed these challenges in the design of Cody, an open-source AI coding engine that incorporates code search and other sources of technical context.
This session will intersperse demos of use cases where Cody eliminates developer toil with technical explanation of how features work under the hood.
We’ll cover common use cases and workflows we’ve observed among the many devs who are using Cody at enterprises ranging from AI startups and driverless taxis to large online marketplaces, 100-year-old banks, and government agencies.
This session will be recorded
