The flagship

Learn AI — From Absolute Basics to Advanced

Go from "what even is a prompt" to shipping AI tools people pay for.

A complete path through modern AI for someone starting from zero. You learn what these models actually are, how to direct them, how to connect them to your own data, and how to build and ship a working AI product — without a computer-science degree and without pretending the math is optional when it is not.

  • Start from zero — no coding background assumed
  • Build 6 real AI tools, not 6 tutorials you watch
  • Ends with a deployed product and a portfolio page
Length
12 weeks
Lessons
84
Projects
6
Language
English

Recorded lessons · downloadable workbooks · live doubt sessions

Opening soon

$19$199

Founding price · lifetime access

Join the founding list

Nothing is charged today. The lessons for Learn AI are still being recorded — join and you keep this price and get first access when it opens.

  • 84 recorded lessons, lifetime access
  • 6 deployed projects with full source code
  • Downloadable workbook for every module
  • Prompt library you keep and reuse
  • Live doubt-clearing session every Saturday
  • Private learner community
  • Certificate on completing all six projects — not on watching the videos
  • Every future update to this course, free

Price includes tax, in US dollars. Paid securely — we never see your card.

Rises to $59 on 1 October 2026 · 4 days away

This is for you if

  • You have used ChatGPT and want to understand what is actually happening.
  • You are a student or working professional who can give this 5 hours a week.
  • You want to build things with AI, not just talk about AI.
  • You have been told "learn AI" for two years and never found a starting point.

This is not for you if

  • You want a certificate to put on LinkedIn without doing the projects.
  • You are already fine-tuning models and reading papers — this will be slow for you.
  • You expect to skip the math week. It is one week and it is load-bearing.

Outcomes

What you will be able to do

Written as capabilities, not results. Every line is something you will have done at least once by the end.

1

Explain in plain language how a large language model produces its next word, and why that explains most of its failures.

2

Write prompts that hold up in production — with structure, examples, constraints and a fallback when the model refuses.

3

Build a retrieval system over your own documents so the model answers from your data instead of guessing.

4

Call the Claude, Gemini and OpenAI APIs from your own code, and choose between them on cost and capability rather than on hype.

5

Design an agent that uses tools, and recognize the three ways agents usually break before you ship one.

6

Fine-tune a small model, and — more usefully — know the four cheaper things to try before you do.

7

Deploy a working AI application to a public URL that other people can use.

8

Read a new model announcement and work out in ten minutes whether it changes anything for you.

Why this course and not a playlist

Built for someone starting at zero

Week one does not assume you know what an API is. It also does not stay there — by week six you are writing code that calls one. The ramp is the whole design.

Six things you build, not six things you watch

A summarizer, a document Q&A tool, a customer-support agent, an image pipeline, a fine-tuned classifier and one product of your own choosing. Each one is deployed, not left in a notebook.

Costed in dollars, honestly

Every project has a running cost, and we show it. Which models are free to use, where the free tiers actually end, and how to build something you can afford to leave switched on.

The failure modes, taught deliberately

Hallucination, prompt injection, context limits, silent truncation, cost blowups. You will meet all five on purpose, in a controlled project, rather than for the first time in front of a customer.

Plain English, jargon explained once

No computer-science background assumed and no term used before it has been defined in ordinary words. Embeddings, tokens, context windows, fine-tuning — each one gets a plain explanation the first time it appears, and then it is simply used.

Updated when the field moves

AI changes every few months. Lessons that go stale get re-recorded and you get them at no extra cost, for as long as the course exists.

The full syllabus

12 modules, 76 lessons

Every module ends in something you physically have. Open any one to read its lessons.

Tools you will use

  • Claude
  • Gemini
  • ChatGPT
  • Python
  • Next.js
  • Supabase + pgvector
  • Hugging Face
  • LangChain
  • Ollama
  • ComfyUI
  • ElevenLabs
  • Cursor

Where a tool costs money, the course says so and shows the free route through the same task.

What this course does not promise

This block is on every course page and it is never removed to make the page read better. It is what makes everything above it believable.

  • This will not make you an AI researcher. It makes you a competent AI builder, which is a different and more employable thing.
  • It does not promise you a job, a client or an income. It gives you six shipped projects and the ability to talk about them.
  • Five hours a week for twelve weeks is the real cost. Watching without building produces nothing.
  • Some lessons will go out of date. We re-record them; we do not pretend they did not.

Questions about this course