Azure Cognitive Services and Custom Vision

If you are dealing with a well-defined general ML problem, such as classification, object or face detection in computer vision, Optical Character Recognition (OCR) and handwriting recognition, speech-to-text and text-to-speech, translation, spell-checking, key word and entity extraction, or sentiment analysis, the chances are high that these services have already been implemented and battle-tested in Azure. In a lot of cases, it greatly saves you time, resources, and effort by reusing these services instead of training similar models from scratch.

If your problem space is very general—such as detecting and matching faces from a camera image to an ID image—or detecting adult content in user-uploaded media, then you can look into Cognitive Services. The Cognitive Services website features demos for almost all the APIs and you can go and try them out for your use case.

If your domain is very specific but uses one of the previously...