Overview
Discover how to leverage large language models for rapid code refactoring and maximize developer productivity in this 25-minute MAAS Show And Tell presentation. Explore the challenges of refactoring legacy code and learn about the LLM code refactoring loop. Gain insights into prompt engineering techniques for optimal results, including crafting effective prompts and improving them through iteration. Examine a real-world example of using LLMs to migrate hundreds of tests from enzyme to @testing-library/react. Delve into automating refactoring at scale with scripting and integrating LLMs into the development workflow. Follow along as Peter Makowski, Senior Web Engineer at Canonical, shares valuable insights and practical strategies for boosting team effectiveness through LLM-assisted coding.
Syllabus
Introduction
Challenges of refactoring large portions of legacy code
LLM code refactoring loop
Prompt engineering for optimal LLM results
Final Prompt
Automating refactoring at scale with scripting
Integrating LLMs into the development workflow
Taught by
Canonical Ubuntu