Azure Virtual Machines Reviews & Insights

I used Azure Virtual Machines in my last organization for deploying out Machine Learning model and related workloads on virtual machines. Our requirement was to enable automated deployment of our compute engine - Databricks, our ML models, and Airflow workflows on scalable virtual machines and Azure Virtual Machines was our choice in the last organization I worked with.

The VM deployment process is really simple in Azure Virtual Machines. But as I said earlier, compute types were a bit limited when I used it. In a few scenarios we had requirements for a Bare Metal GPU instance for high performance compute, but it wasn't available, so we had to look for alternatives.

Azure Virtual Machines serve the use case of when we need virtual machines that are running on the Microsoft Azure Cloud. We are a multi-cloud provider and therefore it's critical for our business value proposition to offer our services on multiple major cloud versions. When you are in the Azure cloud already, Azure Virtual Machines are the only sensible choice of virtual machines.

I would recommend Azure Virtual Machines to anyone who has a business case for building their solutions in Azure. There are reasonable arguments to be made in favor of more stateless architectures for certain use cases, but if you are planning to use virtual machines and are in Azure this is your only option. There are different reasons why one might have a business use case for building in Azure. There are various technical reasons for choosing VMs.