What is a Machine Learning Engineer in this World?

What is a Machine Learning Engineer in this World?

It's not a Software Engineer
It's not a Data Scientist
It's not an MLOps Engineer
It's not a Data Engineer
It's not a Cloud Engineer / DevOps
And yet it's a little bit of all those things at once.

It's being torn between the flexibility needed to prototype and the constraints to impose to smoothly deploy and maintain models.

It's spending the morning pulling your hair out doing Docker in Docker, the afternoon doing transformations on dbt, and the evening to "relax" reading a little scientific publication on the latest advances in LLMs.

It's waking up the next day, looking at your code from the day before and telling yourself it can be simpler, better structured and much better optimized for computing. During the coffee break, learning that what obsesses your colleague ML Engineer himself is learning to optimize CUDA to run a large model on a smartphone.

It's wondering if TensorFlow is still alive or if PyTorch has definitely buried it.

It's having a few strokes of genius, being proud of your CI/CD and a few minutes later miserably failing to cleanly format a yaml file.

It's turning off an EC2/SageMaker instance on AWS when leaving work and on the way home, wondering if it's still on.

It's explaining to your product owner that it's not that simple to prevent an LLM from hallucinating 100% of the time.

It's looking at the code implementation of a library on GitHub rather than its documentation because sometimes it's clearer.

It's spending more time on Udemy or Coursera than in front of your TV.

It's realizing there are no 2 similar profiles of ML Engineers. It's so true that you could be the messiah for some companies and miserably fail the pre-selections for others.

It's seeing your model, once deployed, solve real problems and knowing your work has impact.

It's collaborating with brilliant and interdisciplinary teams.

It's also never stopping learning.

It's realizing you're part of a funny species.

That's what a ML Engineer is like.