I built Kiruko because I was tired of AI explanations that stop at "it uses neural networks." I wanted to understand what actually happens when I type something into ChatGPT. Not the marketing version. The real version. Tokenizers splitting my words into subword units via BPE. Attention heads computing relevance scores across every position. 96 transformer layers processing in sequence. A softmax function sampling from 100,000 vocabulary tokens to pick the next word. So I built an app that teaches all of it. Visually. Interactively. With exams where you have to explain concepts out loud (the app uses AI to grade your spoken answers). 10 modules covering: AI fundamentals, neural networks, self-attention, transformers, training pipelines, LLMs, and a brand-new Generative AI module (tokenization deep-dive, context windows, hallucinations, RLHF, RAG/vector search). It's completely free. No paywall, no ads, no premium tier. I just wanted this to exist. Would love your feedback. What AI concepts would you want explained next?

Kiruko

Learn what happens between your prompt and AI's answer

Learn what happens between your prompt and AI's answer