5 years ago

Category: Deployment, Data, Docker

Suppose a model is running in production and creates predictions that are not understood. Then troubleshooting is needed and at least the following shall be traced back: In this blog, an Azure data science project is introduced using popular Azure services.