genai-learning-os

Overview

GenAI Learning OS is a Chinese-language learning platform for generative AI. It presents a path from mechanism understanding to reliable decisions and uses Kimi K3 real parameters as a starting point to cover architecture, VRAM, inference, API, and evaluation. The site organizes entry points by role, including strategy product/operations, AI application product, R&D/architecture, and complete beginners. It states that role selection does not change course content, only the first step shown, and is saved locally. The curriculum is chapter-based, with a mainline covering Algorithm Basics, AI Decision Mathematics, LLM Systems, Agent Systems, Agent Book, and Model Distillation, plus a parallel image generation track. Learning is structured around a four-step loop: browse, hand-calculate, enter experiment, and review. Case Academy teaches AI through cases, and Agent Book covers Context Engineering, Harness, evaluation criteria, and continuous evolution. The site notes that all formulas, estimates, and experiment results are mechanism-level teaching simulations, not commercial model implementations; production judgment should rely on target models, real logs, and stress tests.