What you'll learn

Large engineering changes become risky when an AI agent edits before understanding the codebase or when recurring automation acts without reviewable evidence. In this course, OpenAI Codex at Scale, you’ll gain the ability to structure Codex work so refactors, migrations, and team automations stay reviewable and verifiable. First, you’ll explore Plan mode, refactoring prompts, ExecPlan records, behavior contracts, and diff review for safe multi-pass cleanup. Next, you’ll discover how to inventory a legacy stack, design compatibility checkpoints, apply framework skills, and validate each migration milestone. Finally, you’ll learn how to run recurring bug triage across Sentry, Slack, Linear, and GitHub, then inspect and recover automation changes through Codex review controls. When you’re finished with this course, you’ll have the skills and knowledge of reviewable Codex engineering workflows needed to scale agent-assisted development beyond one-off coding tasks with stronger planning, validation, and human control.