How to deploy a DeepSeek model on Azure AI foundry as a serverless service and access it privately using a private Endpoint, i need clarification for the DNS part for the mlworkspace for the private endpoint

Question

Hello team,

How to deploy a DeepSeek model on Azure AI foundry as a serverless service and access it privately using a private Endpoint, i need clarification for the DNS part for the mlworkspace for the private endpoint .

what i did is i deployed the model and i created the private endpoint to access it, now how can i inference the model, i need support regarding the part of DNS. Thank you

Answer 1

Hi Ahmed Saber Elkhouly(Administrator),

Thanks for posting your question on Microsoft Q&A.

To deploy a DeepSeek model on Azure AI Foundry as a serverless service and access it privately via a private endpoint, follow these steps:

1. Set Up Private Endpoint & DNS:
   1. Create a private endpoint to securely access the model within your virtual network.
   2. Configure a private DNS zone and link it to your virtual network to ensure proper DNS resolution to private IPs.
   2. Model Inference:
   1. Install the inference package: pip install azure-ai-inference
   2. Use the endpoint URL in the format: https://your-host-name.your-azure-region.inference.ai.azure.com
   3. Authenticate using either a key or Microsoft Entra ID credentials.

For more details, refer to these Microsoft docs:

• Using DeepSeek-R1 with Azure AI Foundry
• Private Link and DNS Integration

If the reply was helpful please don't forget to upvote and/or accept as answer or let me know if you have any other query.

Thank you