Semantic Kernel Workshop

A hands-on workshop exploring Microsoft's Semantic Kernel framework for building intelligent AI applications. This workshop provides practical experience with real-world AI application patterns using Python and Azure OpenAI.

Workshop Overview

This workshop takes you from foundational concepts to advanced implementation patterns through a series of Jupyter notebooks and practical examples. You'll learn how to:

• Build AI applications using Microsoft's Semantic Kernel framework
• Create and orchestrate AI agents with different capabilities and roles
• Construct structured AI workflows using the Process Framework
• Implement enterprise-ready AI features with security and scalability in mind

Prerequisites

• Python 3.9 or higher
• Azure OpenAI API access (API key, endpoint, and deployment name)
• Basic knowledge of Python programming
• Understanding of prompt engineering concepts (helpful but not required)

Workshop Modules

01. Introduction to Semantic Kernel

Learn the fundamentals of Semantic Kernel:

• Core architectural components (Kernel, AI Services, Plugins)
• Building semantic functions with prompts
• Creating native functions with Python code
• Implementing memory for persistent context
• Enabling automatic function calling for AI agents

Key Notebooks:

• 01-intro.ipynb: Core concepts, services, and function creation
• 02-memory.ipynb: Implementing semantic memory with embeddings

02. Semantic Kernel Agents

Master the creation and orchestration of AI agents:

• Creating specialized agents with different personas
• Implementing multi-agent communication patterns
• Agent selection strategies and orchestration
• Building agent topologies for complex scenarios
• Integrating plugins with agents for enhanced capabilities

Key Notebooks:

• 02.1-agents.ipynb: Creating and configuring agents
• 02.2-agents-chats.ipynb: Inter-agent communication and complex patterns

03. Process Framework

Learn to build structured, event-driven AI workflows:

• Understanding the Process Framework architecture
• Defining events, steps, and state management
• Building conversational AI systems with processes
• Implementing complex business logic with AI capabilities
• Creating maintainable and testable AI workflows

Key Notebooks:

• 03.1-intro-to-processes.ipynb: Building stateful, event-driven AI processes

Interactive Playground

Apply what you've learned in a real application:

• A full-stack application with React frontend and FastAPI backend
• Semantic memory implementation
• AI-powered functions and plugins
• Translation and summarization capabilities
• Example of native plugin integration (Weather)

Project Structure

semantic-kernel-workshop/
├── 01-intro-to-semantic-kernel/    # Introduction to core concepts
│   ├── 01-intro.ipynb              # Basic concepts and functions
│   └── 02-memory.ipynb             # Memory implementation
├── 02-semantic-kernel-agents/      # Agent creation and orchestration
│   ├── 02.1-agents.ipynb           # Agent fundamentals
│   ├── 02.2-agents-chats.ipynb     # Multi-agent communication
│   └── .env.sample                 # Environment variables template
├── 03-process-framework/           # Structured AI workflows
│   └── 03.1-intro-to-processes.ipynb  # Process fundamentals
└── playground/                     # Interactive application
    ├── backend/                    # FastAPI server
    ├── frontend/                   # React application
    ├── start.sh                    # Launch script
    └── README.md                   # Playground documentation

Getting Started

1. Clone this repository

2. Create a virtual environment:

   Linux/macOS:

   # Create a virtual environment
   python -m venv venv
   
   # Activate the virtual environment
   source venv/bin/activate

   Windows:

   # Create a virtual environment
   python -m venv venv
   
   # Activate the virtual environment
   venv\Scripts\activate

3. Copy the environment variables template:

   cp .env.example .env

4. Add your Azure OpenAI credentials to the .env file:

   AZURE_OPENAI_DEPLOYMENT=your-deployment-name
   AZURE_OPENAI_API_KEY=your-api-key 
   AZURE_OPENAI_ENDPOINT=your-azure-endpoint
   AZURE_OPENAI_API_VERSION=2024-02-15-preview
   AZURE_OPENAI_EMBEDDING_DEPLOYMENT=your-embedding-deployment
   

5. Start with the first notebook:

   • Begin with 01-intro-to-semantic-kernel/01-intro.ipynb, which includes instructions for installing Semantic Kernel and other required packages.

Learning Path

For optimal learning, follow this progression:

1. Foundational Concepts: Start with 01-intro.ipynb to understand the core components
2. Memory Implementation: Explore 02-memory.ipynb to see how semantic memory works
3. Basic Agents: Create your first agents in 02.1-agents.ipynb
4. Advanced Agent Patterns: Study complex agent interactions in 02.2-agents-chats.ipynb
5. Process Framework: Learn structured workflows with 03.1-intro-to-processes.ipynb
6. Practical Application: Apply everything in the interactive playground

Advanced Topics and Resources

For advanced patterns and enterprise deployment scenarios, explore the Semantic Kernel Advanced Usage repository, which includes:

• Dapr integration for scalable, distributed systems
• Authentication and security patterns
• Natural language to SQL conversion
• Copilot Studio integration
• Microsoft Graph API integration
• Production deployment architecture

Additional Resources

• Semantic Kernel Documentation
• Azure OpenAI Service
• Microsoft Copilot Studio

License

This project is licensed under the MIT License - see the LICENSE file for details.