Google Cloud Functions (GCF) can be integrated with Google’s machine learning services to create serverless applications. To use GCF, you write single-purpose functions that respond to cloud events without the need of a server or runtime environment.
To integrate with Machine Learning (ML), first deploy your ML model on AI Platform. Then, trigger a GCF when new data is available for prediction. The function retrieves the data, makes an API call to the deployed model, and stores the prediction result in a database like Firestore.
For example, if using Vision API, upload an image to Cloud Storage bucket which triggers a GCF. This function sends the image to Vision API, receives labels, and saves them in Firestore.
