I’ve recently heard about this new job title, MLOps, the machine learning analogue of DevOps. I was wondering if anyone is actually working in “MLOps.” What do they do exactly? What backgrounds are they from? Is this a fleeting field or will it become huge like DevOps?
I personally feel the concepts are so new that people don't have a solid sense of all that is possible and thus the skills needed. However, I think it is fair to say that it is part dev ops, part engineering, and understanding how machine learning models work at least at a fundamental level. The latter, so you understand why and how they fail, and to be able to explain that.