The RPI workflow (Research, Plan, Implement) - for advanced AI Coding Agent

Dive into the world of Advanced Context Engineering, tackling the challenges of AI-generated "slop" and high rework rates often found when using coding agents on complex or brownfield codebases. We explore the methodology of frequent intentional compaction, utilizing the "Research, Plan, Implement" workflow to manage context windows and maximize LLM output quality. Learn how this plan-first approach enables engineering teams to maintain mental alignment and reliably ship complex, production-ready code while staying out of the "dumb zone".