Ask HN: How do I become a competitive and competent machine learning engineer?

This kind of rides this question: https://news.ycombinator.com/item?id=13232883

I'm finishing up a bachelor's in mathematics with an emphasis in applied math and stats, I have a solid background in Python, I built a couple small software projects, I do some Kaggle completions, but I mostly pull from libraries like Scikit-learn. And I don't have the grades to go to a top masters program.

Would it help to implement a few algorithms like Naive Bayes and Random Forests?

Should I go for something big like train facial recognition software?

If you're on the hiring side, what separates the ML engineers from the data analysts?