Data Science: Yesterday, Today & Tomorrow

image

Data Science: Yesterday, Today & Tomorrow

by Paul-Louis Pröve

On this years Solutions conference in Hamburg I gave a talk about how the role of Data Scientsts has changed over the years and what type of project setups we might see in the not so distant future. In this article I want to briefly summarize my points.

What’s expected of a Data Scientist today?

Data Scientists have always played a hybrid role, combining scientific methematical knowledge with applied data and programming skills. This made complete sense when the role was rather new and we wanted to connect two areas of expertise that we needed for Machine Learning topics.

However, as the field has grown much larger over the years, we kept changing the expected skills of Data Scientists while sticking to the same name convention. Defining what a DS does not only depends on who you ask but also when you ask this person. My answer to this has changed quite a bit over the past 5 years alone.

Today we use this job title for a large group of experts that sometimes do completely different things. This often results in inaccurate communication leading to trouble in the long run. By today I like to segment Data Scientists into 3 different profiles.

DS does still exist on its own but only if we really talk about the hybrid role it represents. In a majority of cases companies don’t look for this joint role anymore but for one of two other personas.

ML Researchers are the type of people behind new and groundbreaking AI models you might hear about in the news. These people have ditched their applied knowledge in order to focus more on the mathematical research side.

ML Engineers on the other hand might be shaky when it comes to the detailed knowledge of some advanced algorithms. However, they know how to use their set of ML-Tools and apply them in a corporate system architecture. They are programmers.

4 out of 5 times a company wants to hire a Data Scientist, they’re actually looking for an ML Engineer. The problem increases because many of their recruiting material focusses onDS or MLR skills opposed to what they’re actually looking for.

That’s why I believe this differentiation matters. Companies literally hire the wrong people because they interview for ML Researchers while actually needing ML Engineers.

to be continued…