Machine Learning Engineering

In the industry I have noticed the term Machine Learning Engineer being used to describe 2 jobs :-

1. The first being equivalent to a Data Scientist where in an ML Engineer is responsible for solving a business problem through machine learning. This person is responsible for training/validation/testing etc of this algorithm and finally deploying the algorithm to production. Once its in production they monitor it, iterate, retrain the model to better its business value. The person is also expected to be good in software engineering, in case for example they need to write data pipelines to faciliate re-training.


2. The second being a software engineer who is responsible for converting model code to production code, ml infra code, data pipelines for models, model obeservability, designing systems which specifically have ML components and similar tasks. The person requires to know about machine learning however does not participate directly in the process of finding the best ML algorithm for the problem, however he can give his insights or suggestions from an engineering stand point.

1. An ML Engineer who only looks at the engineering side, specifically ML infra, ML obeservability, ML and data pipelines.


2. A Dev Ops Engineer but specifically for ML models

Given we have all these definitions, I identify as an ML Engineer of the first and second kind as well, however i lean more towards the second kind. On my website I intend to talk about how production ML systems are engineered. This includes be talking about both however will go more into the depth of the engineering side. I might skip in depth details of the model, however talking about the model, its architecture, training will be an important topic of discussion.

ML Engg Blog

ML Engineering Blog Posts

February 18, 2024 - What and Why Machine Learning Engineering