⚠️ Active Development
This project is under active development and may contain breaking changes. Stability will improve as we approach a stable release.
This repository implements the Azure AI Baseline Reference Architecture to deploy an AI Foundry Landing Zone with managed compute resources. It aligns with best practices for networking, security, and model hosting.
The solution automates:
- AI Foundry Hub and Projects
- Managed VNet and private endpoints
- Compute clusters for model hosting
- Observability with Log Analytics and Application Insights
Azure AI Foundry Reference Architecture
This project builds on the Azure AI Baseline Reference Architecture to help you design and deploy enterprise-grade generative AI solutions. It incorporates networking, security, and authorization best practices, enabling a scalable and secure AI environment.
🔗 Azure AI Baseline Reference Architecture
🔗 Azure AI Foundry Agents Network-Secured Environment
🔗 Private storage configuration
Please review the quickstart templates that demonstrates how to set up Azure AI Foundry with a network-restricted configuration.
🔗 Azure AI Foundry Template - Network Restricted
🔗 Azure AI Foundry Agents Template - Network Restricted
AI Model Deployment in Azure AI Foundry
To explore model deployment options, including serverless models, fine-tuning, and inference endpoints, refer to the official documentation.
🔗 Deploy AI Models in Azure AI Foundry Portal
🔗 Deploy models as serverless APIs
Follow these key steps to successfully deploy Azure AI Foundry:
- Detailed instructions for deploying Azure AI Foundry, including prerequisites, configuration steps, and setup validation.
- How to test and verify the online endpoint hosting a Hugging Face model to ensure successful deployment and connectivity.
After completing testing, ensure to delete any unused Azure resources or remove the entire Resource Group to avoid incurring additional charges.
This project is licensed under the MIT License, granting permission for commercial and non-commercial use with proper attribution.
This demo application is intended solely for educational and demonstration purposes. It is provided "as-is" without any warranties, and users assume all responsibility for its use.