Eugene  Lockman

Eugene Lockman

1624502640

10 Pitfalls &11 Best-Practices for Design Phase of a Machine Learning Application Project

We concentrate on pitfalls you may encounter in the beginning design phase of a Machine Learning project. We detail our best-practices to avoid these pitfalls.

Have you successfully designed, trained, and tested a machine learning application (MLA)? Despite a vetting in the lab, did the MLA not behave acceptably, maybe even fail in production?

If so, read on, and I shall detail, pitfalls my colleagues and I encountered. I then detail best-practices, some of which are solutions we developed to avoid these pitfalls.

Our Approach

Lowering the high cost and burden of machine learning application design, development, and deployment is the methodology domain of DataOps, DevOps, MLOps, GitOps, CloudOps… xOps, where Ops stands for Operations [1].

We believe that you should learn how to crawl, then walk, then run and then maybe fly, in that order.

#devops

10 Pitfalls &11 Best-Practices for Design Phase of a Machine Learning Application Project