I interview for both MLE and data science roles. Here's what I look for:
MLE: Spark or Hadoop (or some ETL experience), Python
DS: Python, Pandas (or some datafame manipulation experience)
R can substitute for Python, but in tech it's hard to get used to full Python workflows & collaborate without some Python exp.
MLE (more programming and design-heavy): OOP -- do you know what a class is? Can you write good abstractions? Can you design basic DB schemas? How do you debug? Can a DS easily understand your Python code? Do you know basic models and how to write baselines?
DS (more stats-heavy): basics (GLMs, trees, parametric vs nonparametric), strengths and weaknesses of different models, dimensionality reduction, ability to "think critically" about data -- how would you "clean" it? What's the right metric for the problem?
Tell me about a machine learning project you worked on. FYI -- one machine learning project is more than enough. I care more about depth in a project rather than you doing 12 different Kaggle projects that lasted a few days each.
What if [x] assumption didn't hold in your project -- how would you modify the algorithm or pipeline? One can always poke some hole or next step in even the best projects. Can we brainstorm solutions to these issues together?
How do you iterate on experiments? How do you share results / get buy-in? Big red flags: you change multiple variables at once, you don’t think carefully about metrics, you don’t keep track of your experiment artifacts, etc.
Do you have actual opinions on technologies or tools? How do you use them? Ex: deep learning classifiers are bad at imbalanced data but you had lots of data, so you fit a boosted tree classifier on features identified by a deep learning model.
What’s the last paper or technical blog post you read that you found interesting? Explain it to me. Tell me what’s really excites you about it. I am a pretty excitable person, so it shouldn't be too much work to get me excited :)
Do you like collaborating with other people, or do you prefer being siloed? Did this conversation energize you, or did you find it annoying?