Provisioning Azure OpenAI service using Pulumi

Learn how to efficiently provision the Azure OpenAI service using Pulumi, including model deployment and testing.

Azure OpenAI Service provides organizations with a powerful platform to incorporate cutting-edge AI functionalities into their operations. Nonetheless, setting up this service can be challenging, particularly for those who are not well-versed in cloud infrastructure. Additionally, when configuring the AI service, it’s essential to include models to make it fully operational.

In this blog post, we will guide you through the process of provisioning the Azure OpenAI service using Pulumi, a modern infrastructure as code tool.

Creating Resource Group

I have a core stack which I am using to build foundational services for my RAG application — The Azure OpenAI , Azure Container Registry including the resource group.

Creating Azure OpenAI Service

We will use azure-native.cognitiveservices API to create the account. More information on this is here

First, we need to create an Account of type AIServices . The code for that is below.