Learning AI from First Principles
A living documentation of learning AI/ML from first principles, starting with RAG (Retrieval-Augmented Generation). Instead of treating AI as a black box, this project breaks down complex concepts into understandable building blocks using Transformers.js to run models directly in the browser.
The Challenge
AI tutorials either throw around jargon without explanation or skip from concept straight to API usage. Everything seems like magic, not math. Traditional learning requires expensive cloud credits, complex setups, and worrying about API rate limits.
Technology
Transformers.js, TypeScript, React, Vite, Netlify, browser-based ML models (no external APIs)
Relevant Links Additional Resources
Curated documentation, external repositories, and supporting materials related to this experimental protocol.
What this project taught me
How to balance depth with delivery, when to build vs buy, and why documentation is as important as code. I'd love to apply the same rigor to your challenges.