The state of data workers

Data Science, Machine Learning, AI […you name it] all of them are terms in raise in the marketing world as the ultimate solution to any problem but when I see job positions I feel like people is confused. They title the job as “Data scientist” when they are really looking for a “machine learning engineer” and viceversa.

Data scientists, ML engenieers and ML researchers. For sure this positions are not new, but sometimes I feel like people confuses them. After this weekend at EWRL I feel like I needed to share my vision on this more than ever.

Even if a lot of times they do work together, for me these are three different positions.

The Data Science is the guy that analyzes the data to get insights from it. They use Machine Learning as one of their tools but is not their only one. Plain statistics and home made models are sometimes enough to get insights.

From there it takes a machine learning engenieer to build a product from those models or data. It must scale, operate inside a pipeline and run comparisons with the new models generated by the data scientists to check if they run better.

EWRL 2016 > At EWRL 2016

When the models made by the data scientists are not enoguh, researchers come to the rescue. They are the ones that create the new algorithms that data scientists might use. They create the tools that later on lot of us will apply in our work.

I do not have the knowledge to create new tools. I might talk to researchers and discuss about my experience but being a heavy user of their tools doesn’t give me the knowledge to create a new algorithm. Neither the knowledge to scale my models to production.

To orchestrate them you might want to know about the three at least enough to be able to coordinate the group. Or at least that’s my view