Questions d'entretiens - Staff data scientist
79
Questions d'entretien pour Staff Data Scientist partagées par les candidatsPrincipales questions d'entretien
regular questions about ml, stats, like bayes theorem, stats tests and general assumption for solving linear regression, MLE/MAP
1 réponses↳
the theory is covered in almost every book or course of math, stats and general ml Moins
Tell me about a time that you had to get the resources / support of a person not directly in your control or functional group.
1 réponses↳
Talked about a time where I had to work with several functional groups to drive a process change and how I overcame roadblocks / conflicting priorities. Moins
Basic sql questions, data science concepts
1 réponses↳
Why is the recruitment process of a well reputed MNC so bad. The interviewe needs to keep waiting. A feed back is must, let it be negative also. Moins
Derive and implement logistic regression.
1 réponses↳
Standard stuff from ML-101. Poor stuff to test understanding of advanced ML science/engineering. Moins
Case study: Sales increased selling price on all items by 5%. There are 10 items, all with different selling prices. Gross revenue was $500,000 before the price increase, with an average selling price of $1. After the price increase Gross revenue was $505,000, with an average selling price of $0.95. Why hasn't the price increase had the desired impact of increasing revenue and average selling price?
1 réponses↳
Average selling price is impacted by volume / the rate of change. So if your price increase wasn't received well on the higher volume items, then average price could go down. Moins
The Interviewer has a list of problems mentioned below. I was asked to pick one of them and code the algorithm 1. There are few points generated which belong to two classes. I was asked to code the logistic regression without using any external library like scikit-learn or scipy. I was to asked to code entirely in numpy 2. There's a question on solving PCA using iterative methods 3. There's a question on solving eigen vectors using iterative methods 4. There's a question on clustering using k-means on a psuedo generated clusters. 5. One more Question is to implement EM algorithm using numpy. During the process of interview and coding lot's of questions related to optimized numpy code. Ex he asked do you what broadcasting is ?. Mostly he tried to make me do vectorized numpy coding
1 réponses↳
I have picked up the logistic regression question. Frankly I have never done coding in numpy(may be in college during btech). I first wrote highly unoptimized code using for loops and using numpy in each iteration. But later it just struck to my mind that vectorized code does the same job and is faster. I could write the forward pass using the loss function and calculated the batch loss. I was also implementing the gradient update rule using SGD but the interviewer said he is running out of time. Moins
90 minute online exam on codesignal before talking to a recruiter. Coding was ridiculously easy, SQL questions slightly harder, and they had some really silly data science multiple choice questions which involved hand calculating the results of linear and logistic regression.
1 réponses↳
Scored fairly well. There were kind of a lot of questions for the allotted time though. Moins
They asked about my past experience and challenges.
1 réponses↳
Shared my experience and Project challenges and end user experience.
Data Science related questions - conceptually and questions from experience
1 réponses↳
Yes I answered all correct