Ask HN: Data Scientist vs. Machine Learning Engineer
What are the major differences between a data scientist and a machine learning engineer?
Differences in education, tools used, functions within a team, etc.
Differences in education, tools used, functions within a team, etc.
In the case of data science and machine learning, my experience is that data scientists are frequently more involved in munging data, warehousing and writing SQL code for analysis, while machine learning engineers write software or models for interacting with the data and making predictions.
Again, I don't think it's a binary distinction, but personally I'd expect a data scientist to be working mostly with databases and data directly and an ML engineer to be developing software and forecasting models.
Practically speaking the education and background for the two is similar in that they're fundamentally not research oriented positions, they're engineering positions, which means they're accessible to someone with software engineering experience but not a significant mathematics or statistics background. On the converse, a machine learning researcher would be significantly more theoretical - still writing code, but mostly working with abstractions and models, not doing most of the heavy lifting implementation wise. A position like this will be significantly less accessible to, say, most of Hacker News, as you'll generally need an MSc at minimum to be a competitive candidate. Although obviously there are exceptions.
As a piece of advice, I would urge you to consider most of these specialized titles to be marketing terms. If you have an opportunity to work in either role, make sure you're very clear about the actual responsibilities involved. It's a lot easier to throw someone an impressive title than it is to throw them impressive responsibilities or an impressive salary.
But in practice, they're probably not too different from each other. They both need to get data, clean data, analyse data, produce a model that predicts.
In addition, sharing datasets is probably something they have to do: https://blog.helmspoint.com/posts/2017/09/05/sharing-machine...