Data Scientist Interview tips
There are at least three types of data scientists, which usually cause some confusion when preparing for a data scientist interview. Thus, I will firstly classify a little bit.
- data scientist, analytic
Also called data analyzer sometimes focuses on data analysis and visualization, ab test, etc.
what you need to know:
ab test, SQL, business insight, and data visualization, including excel, tableau, redash, etc.
how to prepare:
2. data scientist, engineering
focus on statistics and machine learning (ML), depends on companies, some focus on statistics and some focus on machine learning, and some are more related with the product side, and some are supposed to do more research.
what you need to know:
a deep understanding of ML and statistics, e.g., typical distribution and tests, regression, Maximum likelihood estimation (MLE), Maximum A Posterior (MAP), SVM, EM, K-means, KNN, neural network, decision tree, recommendation system, matrix factorization, ensemble method, deep understanding of CV or NLP or time series, you can read research papers and do experiments
how to prepare:
3. machine-learning engineer
It focuses on model development, deployment, and ML + SDE. You need to code as well as an SDE and have knowledge of machine learning. In a typical interview, you are supposed to tackle medium or hard-level leetcode style algorithm problems and show your understanding of ML. ML design is also essential, depending on what type of role you are applying for. ML design involves data preprocessing, model training, model saving, deploying, online and offline metrics, feedback, and CI/CD.
what you need to know: an end to end ML design, data structure, and algorithms
how to prepare:
And of course, some position is a combination of all three types.
