Hello-Agents is a free, open-source learning project from the Datawhale community that teaches how to design and build modern AI agent systems from the ground up. The curriculum covers agent concepts and history, large language model foundations, classic patterns such as ReAct and reflection, low-code platforms, popular agent frameworks, and the implementation of a custom agent framework using native model APIs.

The material goes beyond basic agent demos to cover context engineering, memory, protocols, evaluation, agentic reinforcement learning, and complete example applications. It includes online reading, local materials, community contributions, and practical exercises intended to move learners from using LLMs to understanding and constructing AI-native agents themselves. Hello-Agents is especially useful for developers and students who want a systematic, hands-on introduction to agent engineering rather than fragmented tutorials focused on one specific framework.