There are a lot of issues taking this course in 2025. Creating a ML Studio instance doesn't grant sufficient permissions for the AI service principle to create container instances in the Azure resource group (have to do that outside of the instructions). Dataset for the first exercise will not render and had to get it from a github repository. Creating a Model-as-a-Service only works for the first exercise (after you find the dataset) which is autoML based (and its called a web service deployment instead). The remaining exercises being all custom-created in designer do not allow deployments through web service, they require real-time or batch deployments that require completely different steps to setup (and in my case, none of the available compute SKUs were available in the region I was in, so cannot be setup at all). You should be able to learn a few things related to training the different model types but not really being able to test/validate them with deployments.

The training is outdated. Very hard to follow along. Spent a lot of time debugging.

I'm facing some issues with lessons and there is zero support to fix it.

outdated instructions and many links not working

The worst course I have ever taken. 5% of the modules work due to updates. Tests have questions not covered by the content. They never tell why you are picking some settings. I have read similar reviews that are 2 years old, seems like coursera doesn't care.

This course is horrible. Instructions for AzureML are not consistent with the user interface, and labs result in errors on the AzureML platform.

The course is out of date. Azure AI (ML studio) interface has changed a lot and it's a struggle to follow through steps . Not worth the time.