This is a part of my deployment series

• 10 Great ML Practices For Python Developers

• Productionizing NLP Models

• 101 For Serving ML Models

• ML Deployment Decision Tree

Besides SOTA models, the second hottest topic in data science is ML infra. There are various tools and many considerations which go into finalizing the pipeline.

The question is not about which is the best tool! It depends on the amount of robustness you are looking for with the constraints of skills and project timeline. Very few companies do this right because of lack of skills and ‘eagerness’ to put models to production for business validation.

I recently did a survey of the tools available and made a decision tree to narrow down on a decently acceptable approach. Although real deployments can be very messy and beyond the possibility of a simple decision tree.

Let me know what do you think. There are a lot of tools nowadays and I might have not covered some.

You can find the tree here - deployment.pratik.ai

Consider upvoting at madewithml.com

Come join Maxpool - A Data Science community to discuss real ML problems!

I am also on Medium, Twitter & LinkedIn.