Introduction

In the previous chapter, we studied model improvements and explored the various techniques within hyperparameter tuning to improve model performance and develop the best model for a given use case. The next step is to deploy the machine learning model into production so that it can be easily consumed by or integrated into a large software product.

Most data science professionals assume that the process of developing machine learning models ends with hyperparameter tuning when we have the best model in place. In reality, the value and impact delivered by a machine learning model is limited (mostly futile) if it isn't deployed and (or) integrated with other software services/products into a large tech ecosystem. Machine learning and software engineering are definitely two separate disciplines. Most data scientists have limited proficiency in understanding the software engineering ecosystem and, similarly, software engineers have a limited understanding of the machine learning field...