Google Data Science interview questions

Overall the process is very smooth and well-explained. First interview is a phone screen with a recruiter who will walk you through Google's interview process. Next step is a 4-part technical screen lasting an hour: 1) Describe a past project 2) Stats question 3) Causal inference question 4) Coding question After that it would have been onsite interviews but I didn't get to that stage. Generally speaking, the types of questions asked were not only more difficult, but also it felt as if the bar was also higher. Positive experience overall, but difficult interview.

I applied for the data Scientist position in the Google Cloud Revenue Acceleration Team. I was referred by a friend and the whole process took about 4 weeks. Phase 1: Discussion with the Recruiter I interviewed for the same role one year ago. I didn't get through and the recruiter at that time refused to provide any specific feedback stating that they were not allowed to share. However, when I reapplied, the recruiter was very helpful and basically walked me through an "approximate" feedback of how I did last time and what went wrong. First technical screen (telephonic): I skipped this part. General Cognitive Ability: This interview focused on business case study like questions (2) and lasted 45 minutes. The idea was to 1. Understand the question and the context quickly 2. Come up with a structured approach towards the solution 3. Discuss the questions posed by the interviewer in the middle of the conversation Overall, preparing for the business case interviews helped me for this interview. The interviewer generally focused and asked questions about the technical aspects of the problem and most of the discussion after coming up with the structure was technical in nature (e.g. What are some ways to effectively reduce the dimensionality of the data? Which models you will pick to answer this (classification) problem and why ? Compare the approach with others etc.) Technical Round (Stats and ML, 45 minutes): This was relatively easy I would say. The interviewer asked to code a simple problem related to string manipulation and percentile calculation. On the ML bit, most discussion hovered around MLE concepts, CI and Hypothesis testing basics. Overall, not very difficult and the interviewer was very helpful. Leadership (45 minutes): In this round , interviewer focused on questions related to leadership experiences in the past and asked for examples showcasing initiative, handling conflicts and pushing / advocating for ideas. The questions were very open ended and the interviewer was very helpful. I suggest writing down 10-15 behavioral questions in advance for practice (which I didn't do). Googlyness (45 minutes): This was simply a rehash of Leadership type questions. The interviewer came prepared with 7-8 classic behavior interview questions and asked them one after the other. By the end I was very tired and I suspect this is the interview which DQed me.