DataWhale 'Build Agents From Scratch' Hits 28K Stars — China's Largest ML Education Org Publishes Definitive Agent Tutorial

GitHub·medium signal

DataWhale, China's dominant ML education organization, released hello-agents ('从零开始构建智能体') covering agent fundamentals through multi-agent coordination in hands-on Python notebooks, accumulating 28.1K GitHub stars. The curriculum spans LLM basics, RAG, tool calling, memory architectures, and orchestration patterns — the Chinese parallel to Simon Willison's agentic engineering guide. The 28K-star traction alongside the DataWhale vibe-coding repos (easy-vibe, vibe-vibe at 3.8K each) signals China's developer education ecosystem is producing its own comprehensive agent engineering canon independent of English-language resources.