Over six sessions, participants built a working AI teammate for their own engineering and program work. They wrote the teammate's brief, built and tuned reusable Skills, gave it a knowledge base and a memory that carries across sessions, climbed messy data from raw file to decision-ready dashboard, and designed thought partners to pressure-test real judgment calls before committing. Each participant completed a capstone: a personalized build roadmap applied to a real problem in their own role.