AI Architect Academy
For people ready to direct AI, not just use it.
$199
- A shipped agentic system
- Credit-worthy for homeschool transcripts
- First year of Recipe Book updates included
- 30-day money-back guarantee
Become an AI architect · The future of computer science
The AI curriculum for anyone who wants to direct AI, not just use it.
The future of computer science is not writing code. It's directing AI agents that build, research, automate, and solve problems for you. This is a hands-on, project-based course for homeschooled high school students and anyone who wants to master the most valuable skill of the AI era. Turn your computer into a team of intelligent agents that work while you learn, create, and live.
Hands-on from day one. Built to stay current and produce a system you actually run.
- 10 modules
- Build a real multi-agent system
- Self-paced
- Capstone project: a personal AI team that solves a real problem
The shift
Traditional computer science taught you to code. The future teaches you to direct intelligence.
A teen with a capable AI coding partner can now produce working code far faster than was possible without one.
The high-value skill is no longer typing syntax.
It's knowing what to ask, what to automate, and how to orchestrate agents across your entire computer. We call that directing AI — and a person who does it well is an AI architect.
The old model
Teach a student to type syntax.
The AI architect
Teach a student to direct a system that writes syntax.
The old model
Optimize for memorizing language features.
The AI architect
Optimize for understanding architecture, state, and cost.
The old model
Ship a toy project at the end.
The AI architect
Ship a working, personally useful agentic system.
Outcomes
What you will actually be able to do.
By the end of the course, you have done the following on your own machine, not watched someone else do it, not read about it.
-
01
Run a local large language model.
Install and run an open AI model on your own laptop. Know when to use local AI and when to use the cloud.
-
02
Direct AI coding partners.
Use AI coding assistants as real collaborators. Tell them what to build, check their work, fix their mistakes, and reject bad code.
-
03
Automate research.
Build a research pipeline that finds sources, checks facts, and produces a brief you can actually hand in or hand over.
-
04
Delegate inbox and calendar work.
Set up agents that triage email, draft replies, and prep you for meetings, safely, under a setup you understand.
-
05
Schedule work that runs without you.
Stand up recurring agents that run on a schedule, with basic monitoring and cost caps. Know what to do when one breaks.
-
06
Orchestrate multi-agent systems.
Chain agents together in sequence, in parallel, or as a team with a leader. Pick the right pattern for the job.
Curriculum
10 Progressive Modules: From AI User to AI Architect.
Each module includes reading, hands-on activities, a quiz, and a project checkpoint that builds toward your capstone project.
-
01
10–14 hrs
The Agent Mental Model
Understand how an AI agent actually reasons, and why that changes how you use it.
-
02
14–18 hrs
Your AI Workstation
Claude desktop app, optional local LLM, dev tools, safe defaults for privacy and cost.
-
03
16–20 hrs
AI Coding Partners
Working with AI coding assistants as real collaborators; reading and directing generated code.
- 04 14–18 hrs
-
05
12–16 hrs
Email & Calendar Agents
Inbox triage and calendar delegation, under a security posture.
-
06
17–20 hrs
Automation & Scheduled Tasks
Recurring agents, monitoring, reliability, cost management.
-
07
14–18 hrs
Extending AI (Skills & Plugins)
Build a custom skill; decide when to build vs. install.
- 08 14–18 hrs
-
09
16–20 hrs
Security, Privacy & Responsibility
Prompt injection, key hygiene, local vs. cloud, ethics.
-
10
28–32 hrs
Capstone — Ship Your Agentic System
Charter, architecture, build, 7-day observation, sign-off.
Capstone
The capstone project is not a final exam. It is a working agentic system running on your own machine.
The deliverable
Students design, build, and operate a real multi-agent system that solves a genuine problem in your life. The system must integrate at least three components across research, coding, scheduling, inbox/calendar, or a custom skill you create.
No final exam. Just a working system running on your own machine that proves you're ready for the AI future.
Why this course is different
Built for the real world. Designed for you.
- Homeschool-friendly
- Flexible and self-paced
- Project-based
- Perfect for high school students and lifelong learners alike
- Don't just learn about AI agents. Build and orchestrate them.
The two-layer architecture
Built to stay current without rewriting itself every quarter.
The hardest problem with any AI course is that the tools move faster than the textbook.
AI Architect Academy is designed around a two-layer architecture. We wrote it this way on purpose.
concept
The Core Book — durable.
Concepts that will still be true in three to five years: how agents reason, how context and memory work, the principles of security and privacy, the patterns of orchestration, the tradeoffs in cost and latency, and the ethics of directing powerful tools.
Tool-agnostic. No product names, no version numbers.
recipe
The Recipe Book — fresh.
Tool-specific, step-by-step walkthroughs. "How to install the Claude desktop app." "How to schedule a recurring research agent in Cowork." Every recipe is dated, versioned, and tagged for review.
Refreshed quarterly. Broken recipes are flagged, updated, or retired.
Credit & transcripts
Designed to hold up when you have to document a credit.
AI Architect Academy is built to satisfy 1.0 high-school credit on the Carnegie unit. Hours, rubric, transcript language, and portfolio assembly are all documented for you, not as marketing, but as the real artifacts a homeschool parent (or umbrella school, or college admissions reader) can hand to a stranger.
-
1.0
High-school credit.
120–150 student hours, budgeted per module.
-
15
Pass/fail capstone rubric.
Fifteen criteria, each pointing to a named artifact. Pass/fail rather than vibes-graded. Optional 0–4 score appendix for households that need a letter grade for a transcript.
-
3
Transcript variants.
One-line, descriptive, and portfolio-style transcript language. Copy whichever fits your record-keeping. Pre-assembled portfolio file list for states that prefer artifacts over titles.
Open the full credit documentation → · Open the printable rubric →
Before you buy
Will this run on my computer?
Most of AI Architect Academy runs entirely in the cloud. Your student can complete the course on virtually any modern laptop with broadband internet and admin rights to install software. The course teaches them to direct cloud-based AI agents (like Claude) to do real work — research, email, automation, custom tools, an end-to-end capstone.
One lesson — Module 2, Lesson 2.3 — has heavier hardware requirements. That lesson teaches your student to install and run a local AI model on their own machine. It's an important part of the curriculum, but it is not required to complete the rest of the course. A student whose laptop can't run local models can read the lesson for the concept, skip the install, and continue cloud-only through Module 10 and the capstone.
Cloud-only path (every lesson except 2.3 install)
- Computer: a Mac, Windows, or Linux laptop or desktop sold in the last 4–5 years.
- Operating system: macOS 12 (Monterey) or newer, or Windows 10 / 11 64-bit, or a current Linux distribution.
- Permissions: admin rights to install software on the student's account.
- Internet: broadband.
- Browser: a current version of Chrome, Safari, Edge, or Firefox.
If the machine can run a modern code editor (VS Code) and stay connected to the internet, it can run the cloud-only path of the course.
Full path including the local-model install (Lesson 2.3)
- RAM: 16 GB recommended; 8 GB works with a smaller model (Mistral 7B instead of Llama 3.1 8B).
- Disk space: 15 GB free for a one-time ~5–8 GB model download.
- Everything in the cloud-only path above also applies.
Apple Silicon Macs (M1 / M2 / M3 / M4) and recent Intel or AMD Windows laptops all work well for local models. If your student's machine can't run local, they can still complete the rest of the course.
Not supported (any path)
- Chromebook as primary device. The course uses VS Code and a coding agent that runs locally; ChromeOS cannot install them.
- School-issued laptops without admin rights. The student needs to install software. If you cannot install apps on the machine, it will not work.
- Tablet or iPad as primary device. The course assumes a real laptop or desktop.
Pricing
The capstone course
$199 to enroll.
A Claude Pro subscription (around $20/month, paid directly to Anthropic) is required to use the AI throughout the course. We don't bill you anything beyond the one-time course fee.
Checkout is secured through Stripe. We do not store card data. 30-day no-questions refund. New to AI? AI Builder Lab is sold separately at Next Frontier Builders.
Audience
Who this is for (and who it isn't).
This course is a strong fit for:
- Motivated learners age 15 or older, including ambitious teens, gap-year students, and homeschoolers seeking credit.
- Capable adult self-learners who want to direct AI at the level of a junior practitioner.
- AI Builder Lab alumni ready for the next step.
This course is probably not for:
- Students with no prior exposure to AI at all. Start with AI Builder Lab, then come here.
- Learners looking for a passive video course. Almost every lesson requires the student to build something on their own machine.
Free lessons
Read the first two lessons before you buy.
Not sure yet? Lessons 1.1 and 1.2 are open in your browser, no signup. They teach the agent mental model everything else in the course is built on. Read them, decide.
FAQ
Questions parents actually ask.
What age is this for?
Sixteen and up, generally. A motivated 15-year-old with some prior AI exposure can absolutely do this — nothing in the content is inappropriate for younger teens. The real gate isn't age, it's whether your student can sustain focus and think in systems across a project that runs several months. If they finished AI Builder Lab or AI for Real Life and were hungry for more, they're ready.
What does my student need to already know?
Not as much as you might think. They should be comfortable on a laptop, willing to read instructions carefully, and able to follow a multi-step technical procedure. Some prior AI use helps a lot — AI Builder Lab is the natural on-ramp. No coding background needed. The course introduces coding concepts as they come up, and students learn to direct AI coding partners rather than write code from scratch.
What equipment do we need?
A reasonably modern laptop with broadband internet, admin rights to install software, and macOS 12+/Windows 10 or 11/recent Linux. That covers virtually every Mac or PC sold in the last four or five years. Most of the course runs in the cloud — your student doesn't need a powerful machine to direct AI agents through research, email, automation, and the capstone.
The one place hardware matters more: Module 2, Lesson 2.3 teaches local-AI install. To do that lesson on your own machine, you'll want 16 GB of RAM (8 GB works with a smaller model). If your student's laptop has less than that, they can still complete the rest of the course on cloud — just skip the local install in that lesson.
What won't work for any of the course: Chromebooks, iPads or tablets as the primary device, and school-issued laptops where you can't install apps. (Cloud access still requires installing the editor and the coding agent.)
We've put together a two-minute pre-purchase checklist you can run on your student's actual machine before you buy. Open the pre-purchase checklist
What if the tools change?
Tools will change — fast. We built the course in two layers specifically to handle that. The conceptual material — how agents reason, how to think about cost and security, how to orchestrate work across tools — stays stable; we call that the Core Book, and it's the part that holds its value over time. The tool-specific walkthroughs (exact buttons, exact setup steps) live in the Recipe Book, which refreshes every quarter so the version you're following is always current.
Are there any costs beyond the course price?
Yes — one ongoing subscription, paid directly to Anthropic, not to us. The course assumes your student has an Anthropic Pro subscription (around $20/month). That single subscription covers the Claude desktop app and everything the course teaches inside it: directing coding edits in the Code tab during Module 3, handing AI agents research, inbox, calendar, scheduled, and automation work in the Cowork tab through Modules 4–8, and Claude.ai chat for general use throughout. The course is built around this one subscription on purpose — one bill, no surprise charges, fully usable after the course ends.
Pro is month-to-month. Time to complete the course will vary by student, and you can pause your subscription between modules if your student needs a break.
Optional advanced path. A handful of later lessons (mainly in Modules 6 and 8) introduce the per-token API path for students who want to build custom scripts and scheduled tasks that go beyond the desktop app's built-in features. That path is fully optional, requires a separate API key with its own monthly cap (we suggest starting at $5–$10), and the course teaches how to set the cap before the first request so there are no surprises. Most students will never need it.
Beyond the Pro subscription, the course has no required paid software. The editor we recommend (VS Code) is free, and the optional local AI stack (Ollama) is free.
Is it self-paced or cohort-based?
Self-paced. There's no scheduled class, no cohort, no Zoom call your student has to make. The one place a real person is required is the capstone sign-off — a reviewer (usually a parent, sometimes a mentor or older sibling) reviews the finished system and signs that it works. We give that reviewer a clear protocol so they know what they're signing off on.
What's your refund policy?
Thirty-day, no-questions refund if the course isn't right for your student. We'd rather lose a sale than hold someone who isn't getting value.
Ready to become an AI architect?
Learn the skill that will define the next decade of technology.