Describe the strategy for classifying a movie based on its synopsis using multiple tags.
Sr Data Scientist Interview Questions
3,383 sr data scientist interview questions shared by candidates
Coding : Data manipulation with/without library. Optimization in Deep learning. Statistics (p-value, chi-square etc.) Few machine learning questions. Dimensionality reduction .
4 things you gonna ask a person, when taking over their model
For stats: Focus on brushing up the stats 101, Confidence interval, Z statistics For ML: Practise case study for example if you're taking over someone's work what are some questions you gonna ask that person and then follow up questions on that For coding: More like leetcode. One main question but would have 3 parts to it. For values: Situational Questions on past work and behavior kind of questions Problem Framing: Simple data manipulation using pandas
Be careful they just rescind offers!
home assignment of recommendation system to build. the model should recommend for n of coins and their value in usd, by a data of past transactions.
Given this array create one that indicates both the integer and keeps track of consecutive runs of the same integer value.
The first round was a comprehensive technical interview focused on Exploratory Data Analysis, feature selection, model building, and evaluation, with a strong emphasis on classification metrics. It also included basic to intermediate Python and SQL questions. The second round was more situational and project-focused. The final directorial round involved questions on my current project, and the interviewer shared his screen to me to walk him through some machine learning code, asking me to explain each line of the implementation.
Here I explain the 6 panels of my virtual on-site interview 1) presentation on the results of my take-home exam: I was supposed to have a 30-minute presentation for 6 people starting at 10:00 am. Till 10:04 no one was in the meeting except me, and finally, I started the presentation with 2 people in the room at 10:10 am. I personally found it very disrespectful. The take home exam in general was simple but annoying and time consuming. I was given 2 dataframes, supposed to be the data for two companies, with no further information and two vague questions to answer. After presenting my approach and results the team had 10 minutes to ask questions but they just asked one question; they asked why did you use regression? And I responded that they question statement encouraged a simple solution and considering 4-hour limitation that was the best solution. 2) hiring manager soft skill interview: this was a textbook behavioral interview. Very typical questions and answers like what would you if you have multiple tasks and deadlines or how do you disagree with your manager 3) Domain-specific interview: this was by far the weirdest interview I've ever had. I do not know what the purpose of it was since the interviewer used 80% of the time to talk about himself; what he does in the company and he barely gave me any chance to talk and ask questions. One of the few minutes I got the chance to say something, he literally interrupted me and did not let me finish. I really cannot believe that amount unprofessionalism, and I want to think that was intentional to evaluate my soft skills! 4) general interview, math and stats: this session was supposed to evaluate my problem-solving skills. This session was very smooth with very easy questions that even a middle school kid solve. One of the questions was if I can fly with speed V around the earth with radius R, how long would it take to complete one rotation? 5) Coding: I was asked by the recruiter to choose between spark python or sql and I had chosen python. The interviewer gave me two schemas (consider them as just the column names of two dataframes) and asked 6 questions. The questions were easy and straight forward; from the mean of a column to do a join and then a groupby to do aggregation. I think anyone with a base knowledge in pandas could have solved these questions. I answered all of them and since we had some extra time, I came up with multiple approaches for some of them. 6) Cross team evaluation: the session started with a question about the source of error in prediction models. That was more of a brain-storming session and sharing my thoughts and approaches since there was no correct or wrong answer.
Past experience, overall business impact, individual contributions to previous projects, familiarity with their products, and typical workflows when building an ML model, and interest in mentorship
Viewing 711 - 720 interview questions