Build an AI operating system
Agents, accountable memory and coordinated work — with twelve lessons and a tested offline lab
A researched studio curriculum for self-study and a facilitated pilot. Start with one owned workflow; build a bounded agent, scoped memory, tool contracts and quality checks; then compare coordinated workers. Apply the same discipline to founder operations, creative studios, organizations, family projects and personal learning. The offline lab uses synthetic fixtures. Cloud, provider and production integrations remain learner implementation work.
Learning Objectives
Define a task with ownership, authority, budget and measurable acceptance.
Build scoped, source-backed memory with correction, deletion and export.
Evaluate one agent against coordinated workers, including failed attempts and review time.
Design recoverable studio, organizational and personal workflows.
Deliver a capstone with test evidence, cost assumptions and rollback.
Prerequisites
- One recurring job you own and can measure.
- Node.js 22 or newer for the offline lab; basic JavaScript for adaptation.
- Synthetic data for all exercises.
- Optional later integrations: provider API access, a development database and a controlled budget.
Workshop Agenda
A useful operating system begins with a unit of work somebody will accept. Choose a weekly job with inputs you can inspect, a result you can test, and an owner who can reject it. Build: One task contract and one measurable baseline.
Share this workshop
Study and build at your own pace
Read the full lessons and download the offline lab directly. No signup is required.