Agentic AI Development

Most engineering teams use AI like autocomplete. This training was about the other mode: treating the model as a teammate you delegate real work to — with a job description, context, and verification.

We covered the full agentic stack: setting up Claude Code as a daily driver, writing effective project context, skills and hooks, running parallel agents on real tickets, and — the part teams skip — how to verify and review AI-written code at speed.

By the end of the day every engineer had shipped a working change through an agentic workflow on their own codebase, not a toy repo.