Jedi

Jedi is Patrick Debois’s Belgian advisory practice focused on how AI agents change software development for individuals, teams, and organizations. Through independent advisory work, workshops, and public writing, Debois explores AI-native development and context engineering: how to turn everyday coding knowledge into guidance that agents can use.

Debois’s work builds on his background in DevOps. He coined the term in 2009, organized the first DevOps Days, and co-authored The DevOps Handbook. His current research includes the Context Development Lifecycle (CDLC), a framework for applying software engineering discipline to the specifications and organizational knowledge supplied to coding agents.

The CDLC gives engineering teams a concrete process: generate structured context, evaluate whether agent outputs match intended behavior, distribute context as versioned packages, and observe production behavior to identify gaps and refine guidance. Its emphasis is on the quality, relevance, and freshness of context, rather than simply giving agents larger context windows.