Azure AI Foundry Workshop

📦Prerequisites | 🚀Quick Start | 🤖Overview | 📔Workshop Content | 🧩Project Structure | ❓Support | 🤝Contributing

🤖 Overview

A hands-on workshop that guides you through building intelligent apps and AI agents on top of Azure AI Foundry, with fun examples related to health and dietary advice. You will:

• Learn Azure AI Foundry fundamentals
• Set up authentication and project configuration
• Deploy and test AI models
• Build AI agents (health advisor examples)
• Implement health calculations and dietary planning
• Evaluate agent performance with safety checks

Duration: 2-4 hours
Focus: Hands-on exercises, interactive notebooks, practical examples, end-to-end project

---

📦 Prerequisites

Before starting the workshop, ensure you have:

• Python 3.10 or higher installed
• An active Azure subscription with access to Azure AI Foundry
• Azure CLI installed
• Git installed
• VS Code, GitHub Codespaces, or Jupyter Notebook environment
• Basic Python programming knowledge
• Model deployment and AI Search connection configured in Azure AI Foundry

---

🚀 Quick Start

1. Clone the repo:

   git clone https://github.com/Azure/azure-ai-foundry-workshop.git
   cd azure-ai-foundry-workshop

2. Install uv:

   # Unix/Linux/macOS
   curl -LsSf https://astral.sh/uv/install.sh | sh
   
   # Windows (PowerShell)
   (Invoke-WebRequest -Uri https://astral.sh/uv/install.ps1 -UseBasicParsing).Content | pwsh

3. Create & activate a virtual environment:

   uv venv
   source .venv/bin/activate  # Windows: .venv\Scripts\activate

4. Set up Azure AI Foundry:

   a. Create Project and Deploy Resources:

   1. Go to Azure AI Foundry
   2. Create a new AI Hub and Project using the AI Foundry Wizard
   3. Deploy required models:
      • GPT models(gpt-4o, gpt-4o-mini) for chat/completion (set TPM to max to avoid issues with Agents notebooks)
      • Embedding model for vector search
      • Ensure the model is deployed in Global-Standard or DataZone-Standard
   4. Set up connections:
      • Configure Bing Search connection
      • Configure Azure AI Search connection
   5. Add your user account to the Azure AI Developer role from Azure AI Foundry Management Portal

   b. Configure Environment Variables:

   cp .env.example .env

   Update .env with your Azure AI Foundry values:

   • PROJECT_CONNECTION_STRING: Your project connection string from Azure ML workspace
   • MODEL_DEPLOYMENT_NAME: Your model deployment name
   • EMBEDDING_MODEL_DEPLOYMENT_NAME: Your embedding model deployment name
   • TENANT_ID: Your tenant ID from Azure portal
   • BING_CONNECTION_NAME: Your Bing search connection name
   • SERVERLESS_MODEL_NAME: Your serverless model name

   Note: The model specified in MODEL_DEPLOYMENT_NAME must be supported by Azure AI Agents Service or Assistants API. See supported models for details. For Grounding with Bing Search, you need to use gpt-4o-mini model.

5. Install dependencies:

   # Install core Azure AI SDKs and Jupyter requirements
   uv pip install azure-identity azure-ai-projects azure-ai-inference[opentelemetry] azure-search-documents azure-ai-evaluation azure-monitor-opentelemetry
   
   # Install Jupyter requirements
   uv pip install ipykernel jupyterlab notebook
   
   # Register the kernel with Jupyter
   python -m ipykernel install --user --name=.venv --display-name="Python (.venv)"
   
   # Install additional requirements (optional - for deploying repo or running mkdocs)
   uv pip install -r requirements.txt

   Note: If you encounter kernel errors in VS Code, try:

   1. Select kernel: Click "Select Kernel" > "Python Environments" > "Python (.venv)"
   2. If kernel is not listed, run python -m ipykernel install --user --name=.venv again, or use the "Create New Kernel" wizard in VS Code to create a new Python environment
   3. Reload VS Code if needed

6. Choose your notebook environment:

   Option A: VS Code

   • Install VS Code Python extension
   • Install either:
     • Jupyter extension for .ipynb files
     • Polyglot Notebooks extension for .dib files
   • Open any notebook and select your Python kernel (.venv)

   Option B: GitHub Codespaces

   • Click "Code" > "Create codespace" on the repository
   • Wait for environment setup
   • Notebooks will be ready to run

   Option C: Jupyter Lab/Notebook

   # Install Jupyter if you haven't already
   uv pip install jupyterlab notebook
   
   # Start Jupyter Lab (recommended)
   jupyter lab
   
   # Or start Jupyter Notebook
   jupyter notebook

7. Follow the Learning Path:

   1. Introduction (1-introduction/)

      • 1-authentication.ipynb: Set up your Azure credentials
      • 2-environment_setup.ipynb: Configure your environment
      • 3-quick_start.ipynb: Learn basic operations

   2. Main Workshop (2-notebooks/)

      • Chat Completion & RAG (1-chat_completion/)
      • Agent Development (2-agent_service/)
      • Quality Attributes (3-quality_attributes/)

---

📔 Workshop Learning Path

Follow these notebooks in sequence to complete the workshop:

1. Introduction (1-introduction/)

Notebook	Description
1. Authentication	Set up Azure credentials and access
2. Environment Setup	Configure your development environment
3. Quick Start	Learn basic Azure AI Foundry operations

2. Main Workshop (2-notebooks/)

Topic	Notebooks
Chat Completion & RAG	• Chat Completion & RaG
Agent Development	• Agent Development
Quality Attributes	• Observability & Evaluations

---

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details on how to:

• Submit bug reports and feature requests
• Submit pull requests
• Follow our coding standards
• Participate in code reviews

---

❓ Support

If you need help or have questions:

---