Review on Superwise by Adam Walsh

The fact that I can write code in Python which then gets deployed directly into my model helps me save time from having to do it manually everytime there is an update or improvement made in TensorFlow. Also, it's super easy to connect my machine learning scripts with other python libraries like Apache Airflow/SageMaker etc. There are some minor bugs where sometimes when you try to run a script using Superwise API it fails but doesn't give much explanation as to what went wrong. It could help if they provided more details about error messages received while running the APIs and also how one could debug those errors. Currently only few functions available at Superwise website are supported i.e. Deployment Scripts, Model Monitoring and Data Preparation. If anyone would require writing machine learning applications using Python this tool should be considered but not for deployment purpose because it has many limitations. One use case I have used the same was creating different data pipelines for ML Models.