The average time to finish the entire interview process for a data scientist position at Microsoft is 2-4 weeks.

• Stage 1: Phone screen
  The recruiter reviews the candidate's resume, research experience, past projects, and knowledge of machine learning & Python.
  
• Stage 2: Code test
  The interviewer sends a link to an interactive coding page and asks the candidate to write R code to solve the coding question.
  
• Stage 3: Full day onsite interview
  Onsite there are a series of interviews that last all day, including a lunch interview. There are back-to-back algorithm interviews. The questions are statistical, inference questions, math modeling questions, managing questions, coding questions etc.

• Q 1: The difference between lasso, ridge; introduce some dimension reduction technique
• Q 2: How to explain a deep learning model to customers?
• Q 3: How to explain p-value to customers.
• Q 4: Basic stats and machine learning questions.
• Q 5: Given a time series dataset, how would you predict the future value?
• Q 6: Replace empty cells with the median
• Q 7: What is the ROC curve and the meaning of sensitivity, specificity, confusion matrix
• Q 8: Generate a fair coin from a biased one.
• Q 9: Generate 7 integers with equal probability from a function which returns 1/0 with probability p and (1-p).
• Q 10: Tell me about your most recent project + questions
• Q 11: Find the maximum of sub sequence in an integer list.
• Q 12: Knowledge of A/B testing
• Q 13: Random-forest and lasso
• Q 14: Code a problem in Python, C++ / C#, R, SQL, Hadoop
• Q 15: Lots of statistics, probability, math, computer science, and database related questions.
• Q 16: How to builds ads model, basic algorithms.
• Q 17: Three friends in Seattle told you it's raining. Each has a 1/3 probability of lying. What's the probability that Seattle is raining?
• Q 18: Can you explain the Naive Bayes fundamentals? How did you set the threshold?
• Q 19: Can you explain what MapReduce is and how it works?
• Q 20: How do you detect if a new observation is an outlier? What is bias-variance trade off?
• Q 21: Can you explain SVM to me?
• Q 22: Perform a series of calculations without a calculator and provide your logic behind the steps.
• Q 23: Discuss how to randomly select a sample from a product user population.
• Q 24: How do you implement autocomplete?
• Q 25: Do you know l1 norm or l0 norm

Come to Microsoft to empower every person and every organization on the planet to achieve more.

To help people and businesses throughout the world realize their full potential.