Much has been written and discussed lately about the data scientist “unicorn” mythology where a single employee personifies a statistician, mathematician, computer scientist, DBA, coder, hardware guru, system admin, and C-suite liaison. Recently, a noted data scientist came out publicly to agree to this over zealous skill-set and set off a heated discussion on LinkedIn.
So it was entirely welcome to see another perspective of the unicorn debate: the Machine Learning Skills Pyramid v1.0 (see graphic below). I agree with machine learning blogger Steve Geringer who previously created the Data Science Venn Diagram v2.0. His new creation lays clear focus on the differences between the ML Researcher, ML Engineer, and Data Engineer. These roles are distinct and should be the basis of a data science “team.” Could you find a single person with the skill-sets for each area? Sure, but the likelihood is very low.
The situation reminds me of a local company that has had a position for Data Scientist open for nearly a year. You see it advertised everywhere. I’ve personally been contacted by three outside recruiters as well as the company’s VP of HR for this job. At the Machine Learning Meetups I attend, I’ve overheard other data scientists chatting about this company, “Hey, have you been recruited for the XYZ job yet?” The sad thing is, just about everyone has. The company has become somewhat of a laughing stock because they’re holding out for a unicorn.
The debate rages on!
Sign up for the free insideBIGDATA newsletter.