CognitoRiverDelta Patreon

Courses

Semester-long courses

A 15-week Computer Science course, and a Business School ladder that meets it in a joint studio. They are meant for young entrepreneurs starting out, and for people already at work who want to lead agent teams at whatever system-design skill they have now.

Enrollment is not open yet. First students are planned for 2027.

In this era, useful software and new products are built by a small human team directing a larger team of AI agents. Most courses still teach you to use a tool. This one asks you to lead a team, once a week, for a whole semester, and to leave with work you can defend.

You join a squad of ten students. A teaching agent is the +1 for your wave. A human instructor stands above that. You start the same week as other squads around the world; together you are a wave. You share a commons, a Friday review, and a harvest. You do your own work in your own fork. When you build something others can use, you offer it back by pull request.

Two semester paths, one joint studio.

A Computer Science course with a published 15-week outline, and a Business School ladder that feeds the same capstone. You can read both and see where they meet.

Computer Science · 15 weeks

Building a Tile-Game Engine the Builder Way

You direct agents to build your own engine, layer by layer, on the same seven layers as the reference engine already running here. You ship six play-tested games: puzzle, teaching sim, tower defence, RTS skirmish, 4X, and a capstone of your own. You keep a decision log of every AI suggestion you reject, and why. You talk for two minutes every Friday, and you sit a mock interview at the end.

JavaScript first, in the browser, on modest hardware. The course assumes data structures, one programming course in JavaScript or a similar language, and basic git. It does not assume prior game development. A later companion course in a compiled language is planned, not yet started.

Business and product development

The Business School ladder, into the same studio

You learn to use agents for research, analysis and operations; then product strategy and market research; then business models, pricing and unit economics for AI services; then go-to-market, partnerships and global operations; then leading mixed human-and-agent teams.

Displacement — who gains, who loses, and how a product can create new roles — is a subject, not a footnote. Business students can audit the engine course’s weekly product thread, so both schools arrive at the joint studio able to work together.

The meeting point

The New Product Development studio

Mixed teams of computer science and business students take one global service from research to launch in a semester. Computer science owns the build. Business owns the market, the model and the launch. Both answer for the result. The brief is rewritten each calendar year.

The 15-week spine.

Each week adds one engine layer and pairs it with one product lesson. Three planned rewrites are kept on purpose: students feel the pain of the simple version before they replace it. Midterm is week 8. Week 15 is a ship, a portfolio, and a mock interview.

Week 1

The builder way

Product: agent teams as a cost. A repository, CI, and one test.

Week 2

Grids

Product: who is the player? Square and hex boards. Lights Out, by hand.

Week 3

Worlds

Product: accessibility and modest hardware. One cell API for every shape.

Week 4

Time and determinism

Product: replays as proof. Rewrite 1

Week 5

Movement

Product: one metric that matters. Pathfinding, reach, terrain cost.

Week 6

Traits

Product: an engine others build on. Rewrite 2

Week 7

Effects as data

Product: teaching and impact. Outbreak teaching sim.

Week 8

Recipes · midterm

Product: a one-page brief. Tower defence. Demo and two-minute talks.

Week 9

Economy and territory

Product: pricing and unit economics. RTS skirmish.

Week 10

Rival AI

Product: AI that explains itself. Rewrite 3

Week 11

Progression

Product: retention without manipulation. 4X: Little Planet.

Week 12

Presentation

Product: localisation. Drawing and sound only listen to events.

Week 13

Saves, fog, multiplayer

Product: operating a global service. Replay and spectator mode.

Week 14

Capstone build

Product: go-to-market and a launch checklist. A game of your own.

Week 15

Ship

Product: jobs gained and created. Final demo, portfolio, mock interviews.

On Friday each team summarizes the weekly sprint

Every Friday: a demo and play-test; a two-minute talk, timed; an interview round; one live pairing with AI allowed; and one explanation with no AI. Grades reward tested, explained work, including the decision log of suggestions you overruled.

A portfolio

An engine, six games and play-test reports — or a launched-style product brief from the business path.

The interview, practised

AI allowed in one round, no AI in the next, and stories of overruling a model. A mock interview at the end.

A real squad

Ten students, a +1, a weekly harvest, and work offered back by pull request.

A public method

You can point at the method, and at public example projects, from the same organization that teaches the course.

Proof, not promises.

The course trails a reference engine that already exists. Work that remains private is not used as a public demo.

The reference engine

The shared library crd-shared-worlds already has four iterations done and 126 tests passing. Iteration 5 is the goal for the week of 16 November 2026. Students build on the same seven layers.

Four iterations done. 126 tests passing.

Work already in public

Orbit Lab and the already-public games show the builder way in the open, with a record of how each one was made. Some of our example builder projects — feedback welcome.

See the work →

Honest status: a 2027 start, a small pilot first.

First students in 2027. The first runs are free and not for credit: one Team +1 of ten students, plus one who leads, works through the course together. Fall 2027 is the proposed start of continuous weekly waves. This page does not pretend that applications are open.

A note for interested students

If you are interested in learning how to use a team of AI agents to build solutions, we want to hear from you and learn what courses you are most interested in. We have many ideas but will be narrowing our focus to one course for Computer Science curricula and one course for the Business School, both pertaining to working in tightly focused teams of the 10+1 model of 10 students + 1 leader, the same model the students will each be learning where 10 AI agents will be the team they will lead each day to deliver value to their teams and to the class, following real world software development and project management skills in daily and weekly sprints.

We are not accepting enrollment yet, in these early days, but we’re working on some exciting ideas and hope to hear from you to learn what you may be most interested in learning regarding Agentic AI and the Builder Way of building the CognitoRiverDelta Way!

There is no enrollment form on this page, and we are not collecting names or addresses here. Enrollment is not open yet.