ML Engineer is an engineer who can train, tune, validate, scale, deploy, monitor and maintain a production ML system. ML engineer should be on engineering teams and work closely with data engineers and front end engineers. Work for ML engineers is less open ended and much more well defined like most engineering tasks (this is NOT to imply an ML engineer’s task is easy or doesn’t have room for creativity). Data scientist work is open ended, like that of scientists, focused on finding the right features, the right statistical approach, designing the right error functions to express the optimization goal, hypothesis driven etc. The data scientist is often stretched to also engineer a production ready system that can scale.
Ideally you want to have engineers who understand data science enough to not only understand a data scientist’s work but also able to translate it into a well engineered ML system. That’s scalable, performant, robust and cost efficient.