AI 入门

实战

• 《hello-agents》 从零开始的智能体原理与实践教程
• 《智能体设计模式》

框架

• helloagents A agent framework based on the tutorial hello-agents
• autogen A programming framework for agentic AI
  • 《Ultimate-Agentic-AI-with-AutoGen-for-Enterprise-Automation》
  • AutoGenBench a benchmarking suite for evaluating agent performance.
• ds4 (DwarfStar) antirez 开源的高性能原生推理引擎与编码 Agent（支持 DeepSeek V4 Flash/PRO、GLM 5.2）

Prompt Engineering

• Prompt-Engineering-Guide prompt engineering, context engineering
• openplayground An LLM playground you can run on your laptop

  pip install openplayground
  openplayground run

• LM Studio Local AI, on Your Computer.
• ragas Supercharge Your LLM Application Evaluations.

MCP

• MCP
• fastmcp The fast, Pythonic way to build MCP servers and clients.
• mcp-zero Model Context Protocol (MCP) server for go-zero framework - Generate APIs, RPC services, and models with AI assistance.
• modelcontextprotocol/go-sdk

Test

• promptfoo 极其简单的 CLI 工具，通过 YAML 配置文件就能横向对比不同 Prompt 和模型的表现。
• DeepEval 类似 Pytest 的单元测试框架，非常适合开发者。

视频教程

• 《AI FOR EVERYONE》

• 《机器学习实战：基于 Scikit-Learn、Keras 和 TensorFlow：原书第 2 版》

吴恩达的课程

• 《2022 年吴恩达机器学习课程》

• 课后练习

Go语言

• 《adk-go》 An open-source, code-first Go toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.