Training an ML model takes ~50 mins + 6 gb once a week. I have specific C# code it needs - can't use python/R.
Can't do Azure Functions since it is a 10min/1.5gig limit. Don't want to provision an Azure App Service b/c I only use it once a week - would be expensive for that amount of GB. Can't do a VM b/c it would be a security risk even IF I wanted to set up a start/stop to not get charged, which I don't.
Can a pipeline/build-service handle 50 mins & 6 gb? Or, what would be a better method to run what /would/ be an Azure Function for ~50 minutes, using 6gb on demand?