For many people who get into a PhD program, their goal is to learn and become an expert in a field they care about. It might be to become a professor explicity, or it might be to become a research scientist. It might be improve their lot in life. Whatever it might be, I dont presume to know.

However ...

My Machine Learning Journey

Today, the bulk of my work today is related to applied machine learning - building tools that integrate machine learning models (see handtrack.js a library for handtracking in the browser, neuralqa - a library for neural question answering on large datasets) or training models from scratch to solve specific problems (see data2vis - a model automated genration of visualizations from data or signver - a library for automated signature verification in documents). However, my formal education and training was not in the field of machine learning.

So how did I come to a point where I built enough competence to contribute novel insights to the field?

Resources

• [x] Add experiences from my twitter bookmarks https://twitter.com/vagar112/status/1509356589299339264? s=20&t=EdJFroGMpLs1CnYPEcrPYQ
• [x] Add Chris Molnar tweet on decoupling

Conclusion

There are many reasons why and industry career could be a better career choice. Ofcourse, there are multiple labs and research groups that go over and beyond in creating a great experience and there are many people who go on to enjoy their academic careers. And they are doing important work educating the next generation of scientists. This reflection is not meant to actively discourage a career in academic but more to help others understand aspects of it that may not be immediately obvious, and more importantly - the availability of options.