Getting models into production

The path from building an accurate model in a lab to deploying it in a product involves collaboration of data science and engineering, as shown in the following three steps:

1. Data research and hypothesis building involves modeling the problem and executing initial evaluation.
2. Solution building and implementation is where your model finds its way into the product flow by rewriting it into more efficient, stable, and scalable code.
3. Online evaluation is the last stage where the model is evaluated with live data using A/B testing on business objectives.

This is better illustrated in the following diagram: