This repository contains hands-on labs demonstrating the integration of Azure AI Agents Service with Semantic Kernel. The labs are designed to help developers understand and implement AI-powered solutions using Microsoft's latest AI technologies.
This project was created during a hackathon to showcase practical applications of Azure AI Agents and Semantic Kernel working together. Through a series of Jupyter notebooks and sample applications, you'll learn how to work with both single agents and multi-agent systems.
The repository includes:
- Interactive Jupyter notebooks that teach the fundamentals of Azure AI Agents and Semantic Kernel
- Step-by-step tutorials progressing from basic single-agent scenarios to complex multi-agent interactions
- Sample applications demonstrating practical implementations of concepts covered in the notebooks
- Complete example scenarios showing real-world applications of AI agents
- Azure subscription
- Azure AI Services account
- Python 11.0.0 or later
- Visual Studio Code or Visual Studio 2022
- Basic knowledge of Python and async programming
- Azure CLI installed
-
labs/
: Contains hands-on Jupyter notebooks with interactive tutorialslab_1.ipynb
: Introduction to Semantic Kernel Agents- Basic agent creation and configuration
- Chat history and agent interactions
- Function calling and plugins integration
- Practical exercises with single agents
lab_2.ipynb
: Multi-Agent Systems with Semantic Kernel- Transitioning from single to multi-agent systems
- Agent collaboration using AgentGroupChat
- Specialized agent roles and team design
- Agent selection and termination strategies
-
solutions/
: Practical implementations showcasing concepts from the labsrws-app/
: Complete multi-agent application for infrastructure managementsetup/
: Infrastructure deployment with Azure Bicep templates and configuration files- Infrastructure as Code (IaC) using Bicep for Azure resources
- Azure Functions for backend services (SQL, Weather)
- Configuration for Azure API Management and Azure Cognitive Search
src/
: Source code for the multi-agent system- Specialized agents for infrastructure analysis, water management, and business advising
- RAG plugin for knowledge retrieval from documents
- API plugin for connecting to backend services
- Agent collaboration frameworks for sequential and custom workflows
- Clone this repository
- Configure your Azure credentials
- Follow the lab instructions in each directory
- Exploring the Semantic Kernel ChatCompletionAgent
- Exploring the Semantic Kernel OpenAIAssistantAgent
- Exploring the Semantic Kernel AzureAIAgent
- Exploring Agent Collaboration in AgentChat
- Understanding the kernel
- An Overview of the Agent Architecture
- Semantic Kernel Components
- Function calling with chat completion
This project is licensed under the MIT License - see the LICENSE file for details.