Building an ML driven image tagger on Google Cloud Platform

In this video, we're going to build a simple image tagger that can tag an user uploaded image with keywords based on the objects present in the image. Such a system provides great value for downstream applications like search. In this tutorial, we'll be using 4 GCP (Google Cloud Platform) services; GCS, Cloud vision API, DataStore, App Engine. We'll be using Java + Spring Framework to implement most of the logic. One important detail I left out from the video: DO NOT forget to shut down the GCP project once you're done with it, in order to avoid being charged. Code: https://github.com/thushv89/tutorials_deeplearninghero/tree/master/ml_image_tagger Readme: https://github.com/thushv89/tutorials_deeplearninghero/blob/master/ml_image_tagger/image-server/README.md Additional resources Service account impersonation: https://cloud.google.com/iam/docs/service-account-impersonation 00:00 - Introduction 02:38 - GCP service pricing 04:38 - Why Java 06:08 - Creating GCP resources 16:19 - Setting up Java and Spring 22:21 - Explaining the backend code 29:00 - Testing the application locally 41:30 - Deploying the app to App Engine 44:30 - Testing the deployed service 48:52 - Final thoughts and Conclusion