Motivation

Motivation

Machine Learning (ML) is not accessible to everyone. There are many common tasks where we have figured out the right way to build the models. We don't have to re-invent the wheel for every similar problem. We can leverage the learning and make it simple. Though this is the way industry is moving, there is no good free and opensource solution which is robust and works on both machine learning and deep learning (DL).

We have tools which looks at certain aspects of ML development but not the full cycle. To put it in a short way, a tool which can take a task description and data and gives a deployed model which can be used directly. We need a simple to use tool which can automate machine learning and deep learning training without bothering us with all the details. Make it as simple as a form fill and press of a button to get a model which can be used in production.

The problem today is of the choice. Too many frameworks, technology options for handling data, back-end etc. This makes it develop a solution which can fit into every need. This is where experience comes in. We have to make some decisions on the choices and stick to it so that the solution is usable for lot of use cases. We believe in standardization of interfaces. By doing this we can use common modules with those interfaces to cross use across applications.