Model Deployment - What and Why | AI Planet (formerly DPhi)

Learning Objectives

You have spent several weeks or months building a machine learning model that would identify if a person has put a mask on their face or not with an excellent accuracy score.
Sounds great, right? Is this all you wanted?
No. Building a model is generally not the end of the project.
You would want your model to be used in real-time where it could identify people in parks, at bus stations, streets, etc., with no mask and immediately inform the people nearby to maintain sufficient distance.
This is where model deployment comes into the picture.

The concept of deployment in data science refers to applying the model for prediction using new data.
In other words, model deployment means using your trained ML model to predict new data available to users or other systems.
In technical terms, model deployment means integrating a machine learning model into an existing production environment where it can take in an input and return an output.

A machine learning model can begin to add value to an organization only when that model’s insight is timely available to the users for which it was built.
Even if the model aims to increase the knowledge of the data, the knowledge gained should be organized and presented in a way that a customer will use it.
If you directly present the model’s code, the customers cannot understand. You need to provide a user interface.
To achieve this, you need to deploy the model on the web.