Welcome to Docker DevDuck Multi-Agent Workshop¶
Workshop Duration
Approximately 1-2 hours of hands-on learning
Welcome to a Docker DevDuck Multi-Agent workshop that will guide you through building and deploying a Multi-agent system using Docker, Google Agent Development Kit (ADK), and Cerebras AI.
What You'll Learn¶
In this hands-on workshop, you will:
- Build and deploy a multi-agent system using Agentic Compose
- Understand how to orchestrate multiple AI agents for specialized tasks
- Learn to integrate Cerebras AI with local models for enhanced programming assistance
- Master Docker containerization for AI agent coordination
- Explore FastAPI web interfaces for agent interaction
- Practice with real-world Node.js programming scenarios using AI agents
Workshop Structure¶
This workshop consists of both theoretical concepts and practical exercises designed to give you hands-on experience with:
🎼 Agent Orchestration¶
Learn how Docker Compose coordinates multiple AI agents in a unified system.
🤖 Multi-Agent Communication¶
Understanding inter-agent messaging and intelligent request routing.
💻 Local and Cloud AI Integration¶
Combining local models with Cerebras cloud services for optimal performance.
🌐 Web Interface Development¶
Building user-friendly interfaces for seamless agent interaction.
Real-world Applications¶
Throughout this workshop, you'll work with practical Node.js development assistance scenarios that demonstrate the power of coordinated AI agents in software development:
- Code generation and explanation
- Complex problem-solving for Node.js applications
- Project coordination and development planning
- Code analysis and optimization suggestions
- Testing strategy development
- Educational programming assistance
Ready to Start?
Make sure you have reviewed the prerequisites and have your development environment ready. The next section will guide you through the necessary setup requirements.
GitHub Sources¶
The source code for this workshop is available at https://github.com/dockersamples/docker-cerebras-demo
Run this workshop in a Docker container¶
Clone this repo and run the following command:
Now you can access the workshop at http://localhost:8002