It's been about three months since I'm working as an MLE. The most frequent question I've been asked is that I don't have any ML research, how I am working! This is true that I've been working as an SWE since early 2019 and I have research in a distributed system for the automotive industry and yet not published! What I experienced in these three months is there is a good borderline between MLE and ML researchers. and obviously MLE vs SWE.
Firstly in SWE, everything is practical, you get immediate feedback on everything you do, It may be good or bad ! whereas MLE is extremely ambiguous. The problem might be in data cleaning, hyperparameters tuning, wrong models selection, or even wrong data for training, etc.
Secondly, ML researcher are smart people who can build interesting models, insights, predictions and also use some libraries on the other hand MLE can write maintainable, sustainable code, efficient way to store the data, and deploy the models so that it can be accessed from almost every machine and at the same time should understand the models, basic understanding of the maths behind the model.
These are my experience, correct me if I am wrong. One thing I mentioned first about academic research or advance higher degree. To some extent, I think it will be beneficial to do Msc in applied mathematics or specialization in ML if an experienced MLE wants to pursue more.
Top comments (0)