3 years ago

Category: Software, Privacy

If you are anything like me and have binge-watched all episodes of Shark Tank then firstly, hello new best friend :) and second, more importantly, you are halfway there on how to answer this (tricky) interview question! Say, for example, you are invited to Shark Tank (you lucky duck) to pitch your new restaurant idea and this question comes up, what they want to know is — what are your plans to make this idea a functional idea?

There are several options when it comes to picking a framework (for instance SkLearn, Pytorch, Keras, Tensorflow, H2o, Theano) for your ML and Deep Learning solution and it is important that you make a choice that will serve you well in the long run.

When that happens, make sure that the addition of these features is not leading towards overfitting the model and that you are able to witness an improvement on your validation sets as well.

As your model becomes (or is predicted to become) more and more popular, i.e. the number of prediction requests keep pouring in, it is time to think about whether your model is churning out predictions fast enough.