Content
# 5-Day AI Agents Intensive Course with Google
> **42万+开发者参与的 AI Agent 课程** | 感谢 [sdivyanshu90](https://github.com/sdivyanshu90/5-Day-AI-Agents-Intensive-Course-with-Google) 和 Google 提供的免费课程资源
5 天直播回放视频:[youtube](https://www.youtube.com/playlist?list=PLqFaTIg4myu9r7uRoNfbJhHUbLp-1t1YE)
谷歌新课程:[25天Agents课程-圣诞节礼物](https://mp.weixin.qq.com/s/H4C65Vvh58G3g8AZ1vl1lA),学习笔记:[25-Day-Agents-Course-by-Google](https://github.com/anxiong2025/25-Day-Agents-Course-by-Google)
## Quick Start
### Prerequisites
- Python 3.12+
- [uv](https://docs.astral.sh/uv/) package management tool
- Google API Key ([获取地址](https://aistudio.google.com/apikey))
### Installation and Configuration
```bash
# Install dependencies
uv sync
# Configure API key
cp .env.example .env
# Edit the .env file and add your GOOGLE_API_KEY
```
### VS Code Configuration
**Required Plugin**: Jupyter
- Install: `Cmd+Shift+P` → "Extensions: Install Extensions" → Search "Jupyter"
- Select Kernel: Click the kernel selector in the notebook → Select `.venv (Python 3.12.x)`
### Run Notebook
Open `day1/*.ipynb` in VS Code and press `Shift+Enter` to run the cell
---
## Course Outline and Daily Assignments
### ✅ Day 1: Agent Introduction
**Topic**: Agent Classification, Agent Ops, Interoperability and Security Basics
- 📄 [Whitepaper: Introduction to Agents](https://www.kaggle.com/whitepaper-agents)
- 🎙️ [Podcast: Unit 1 Summary](https://www.kaggle.com/whitepaper-agents-podcast)
- 💻 **Code Experiments**:
- [Build Your First Agent with Gemini and ADK](https://www.kaggle.com/code/markishere/day-1-prompting-with-gemini)
- [Build Your First Multi-Agent System](https://www.kaggle.com/code/markishere/day-1-agent-architectures)
- 📚 [Code Experiment Troubleshooting Guide](https://www.kaggle.com/discussions/general/552193)
**Local Notebooks**:
- `day1/day1-01-from-prompt-to-action.ipynb` - Basic Agent + Google Search
- `day1/day1-02-agent-architectures.ipynb` - Multi-Agent Mode (Sequential/Parallel/Dynamic)
- `day1/day1-01-First-Agent-Web-UI.py` - FastAPI Web Interface (Run: `uv run day1/day1-01-First-Agent-Web-UI.py`)
---
### Day 2: Agent Tools & Interoperability (MCP)
**Topic**: External Tools, Real-Time Data Retrieval, Model Context Protocol (MCP)
- 📄 [Whitepaper: Agent Tools & Interoperability with MCP](https://www.kaggle.com/whitepaper-agents-tools)
- 🎙️ [Podcast: Unit 2 Summary](https://www.kaggle.com/whitepaper-agents-tools-podcast)
- 💻 **Code Experiments**:
- [Extend Agent Capabilities with New Tools](https://www.kaggle.com/code/markishere/day-2-tools)
- [Tool Best Practices: MCP and Long-Running Operations](https://www.kaggle.com/code/markishere/day-2-mcp-and-long-running-tools)
---
### Day 3: Context Engineering: Sessions & Memory
**Topic**: Context Window Management, Sessions, Memory (Long-Term Persistence)
- 📄 [Whitepaper: Context Engineering: Sessions & Memory](https://www.kaggle.com/whitepaper-agents-memory)
- 🎙️ [Podcast: Unit 3 Summary](https://www.kaggle.com/whitepaper-agents-memory-podcast)
- 💻 **Code Experiments**:
- [Implement Session Management for Instant Context](https://www.kaggle.com/code/markishere/day-3-sessions)
- [Implement a Memory System for Long-Term Personalization](https://www.kaggle.com/code/markishere/day-3-memory)
---
### Day 4: Agent Quality Assurance
**Topic**: Evaluation Framework, Observability (Logs/Traces/Metrics), LLM-as-a-Judge, Human-in-the-Loop (HITL)
- 📄 [Whitepaper: Agent Quality](https://www.kaggle.com/whitepaper-agents-quality)
- 🎙️ [Podcast: Unit 4 Summary](https://www.kaggle.com/whitepaper-agents-quality-podcast)
- 💻 **Code Experiments**:
- [Implement Observability for Debugging](https://www.kaggle.com/code/markishere/day-4-observability)
- [Evaluate Your Agent](https://www.kaggle.com/code/markishere/day-4-evaluation)
---
### Day 5: From Prototype to Production
**Topic**: Deployment, Scaling, Agent2Agent (A2A) Protocol, Vertex AI Agent Engine
- 📄 [Whitepaper: Prototype to Production](https://www.kaggle.com/whitepaper-agents-production)
- 🎙️ [Podcast: Unit 5 Summary](https://www.kaggle.com/whitepaper-agents-production-podcast)
- 💻 **Code Experiments**:
- [Implement Multi-Agent Communication Using A2A Protocol](https://www.kaggle.com/code/markishere/day-5-a2a)
- [[Optional] Deploy to Google Cloud Agent Engine](https://www.kaggle.com/code/markishere/day-5-agent-engine)
---
## Notes
- Kaggle code experiments require phone verification
- The project uses `.env` to store API keys (do not commit `.env` to git)
Connection Info
You Might Also Like
markitdown
MarkItDown-MCP is a lightweight server for converting URIs to Markdown.
servers
Model Context Protocol Servers
Time
A Model Context Protocol server for time and timezone conversions.
Filesystem
Node.js MCP Server for filesystem operations with dynamic access control.
Sequential Thinking
A structured MCP server for dynamic problem-solving and reflective thinking.
git
A Model Context Protocol server for Git automation and interaction.