NITIN MUKESH

First Edition

© 2021 Nitin Mukesh

All rights reserved. No portion of this book may be

reproduced in any form without permission from the
Author.
TABLE OF CONTENTS
I. INTERVIEW PREPARATION
1.1 How data science interview is different
1.2 Data science campus recruitment Process
1.3 Prepare Aptitude effectively
1.4 Prepare for Data Science Interview
1.5 Is that all asked?
1.6 Some tips for interview preparation

2. INTERVIEW QUESTIONS
2.1 Statistics interview Question
2.2 Regression analysis Questions
2.3 Logistic regression Questions
2.4 Machine learning Questions
2.5 Coding Questions
2.6 Guestiates Questions
2.7 Puzzle Questions
2.8 HR Questions.

3. MISCELLANEOUS
3.1 Tell me about yourself.
3.2 How to be motivated during
placement season?
Clearing a data science interview isn't a big deal if you
are well prepared. But the fact is that data science
borrows concepts from many fields and that's why you
need to have knowledge of all those fields to clear a
data science interview.

Though there are different jobs related to data like data

analyst, data engineer, Big data developer, all these
domains are not the same.

Data scientist's job profiles differ from all these in a lot

of ways.

There are a lot of concepts that you need to be aware of

for clearing a data science interview.

Believe me, I have gone through a lot of interviews and

failed many. The reason was that, if you lack in any of
the areas, you will be discarded.

But, this doesn't make data science job interviews so

complicated. What is required for a data science job, is
already with you.

You just need to sharpen those skills and along with

them, learn some new concepts.
This ebook is not for a detailed explanation of these
topics, but it is about what to read, where to read
from, and how to prepare for interview. Nitin Mukesh
If I want to summarize the
entire campus recruitment
process for data science
jobs, it can be divided into
two sections:

Technical Tests
Interviews

In campus recruitment, or in off-campus jobs too, the

first stage is to appear for a test. Those students who
clear the test are asked to appear for an interview.

In the data science domain, the technical tests are

mostly centered around common aptitude sections
which include, quantitative aptitude, reasoning,
sometimes English, or a few times coding.

Different companies have different types of tests. For

these tests, you should prepare basic aptitude.

Coding tests are there in few companies only.

From a lot of tests that I have appeared in or seen,

during my campus placement journey, here are few
examples of which companies take what type of tests
for their data science profile:
Nitin Mukesh
 Aptitude: Citi bank, ANZ, Wells Fargo, ICICI
Lombard, IDFC First Bank, PharmaACE, Accenture,
FCS Limited, General Mills,
Aptitude+Coding: Delhivery, Goldman Sachs,
Miko.ai, Flipkart, Icertis, Postman, TCS Research,
Honda R&D,
Data Science-based test: Microsoft, Micron
Technology

Now, as you can see, most of the companies go for

aptitude tests. Though the companies taking coding
rounds visit some of the premier campuses only.

You should find out which company visits your college

each year and what are their recruitment processes.

This is a really nice strategy to target a company that

you want to be in.

You should always do your own research during the

campus recruitment drive.
Nitin Mukesh
To prepare for aptitude sections, you can use the
following resources:

Quantitative aptitude by Dr. R. S. Agarwaal

A Modern Approach To Verbal & Non-Verbal
Reasoning By R.S. Aggarwal
Quantitative Aptitude by Arun Sharma
https://www.indiabix.com/
https://www.pariksha.co/placements

For beginners, Quantitative aptitude by R S Agarwal is

really a good book. I have personally used this book
for my aptitude preparations.
Most of the students feel comfortable with the
aptitude section. Mostly aptitude can be solved
without extra preparations.

But to get accuracy and speed, you can think of

preparing for it too.

If you are thinking of preparing for aptitude from any

book mentioned above, it can take a lot of time if you
go solving each and every question.

I mean the Book By R. S. Agarwaal has more than 700

pages,
Nitin Mukesh
Reading the whole book will require a lot of time. Also,
you need to prepare for the interview part. So, it is not
possible to dedicate most of the preparation time to
the aptitude section only.

But, there is a strategy that you can opt.

10-2-4 Rule

What I call it is the 10-2-4 Rule.

So according to this rule, you have to follow the given

steps each day:
Read the formula given at beginning of a chapter,
solve 10 questions from that chapter.
From the previous two chapters solve 4 questions
from each.

This way you will be able to go through all the chapters

in 15-20 days. Now, each day solve few questions from
4-5 chapters at once.

Also, appear for the online mock test that you can get
on the website given above.

Even if you feel confident, keep practicing few

questions on regular basis. Nitin Mukesh
What I have personally experienced is that in the
beginning, we prepare aptitude very well, but as the
campus placement goes on, if we are not selected in
any company or any interview, we leave practicing
aptitude and only focuses on interview.

Just solve a few questions. Maybe once a week you can

do it so that in the next company you will be able to
perform well again.

Enough about the aptitude, lets now move to the

interview preparation section.

Now if you have cleared the aptitude round, here

comes the interview. The interview questions are
mainly centered on your resume, job profile, and your
academic course(if you are from mathematics,
statistics, computer science, or related fields).

But few topics are commonly asked if you are

appearing for a data science interview. You should
brush up on these basics concepts and also try to have
some mock interviews in prior.

Nitin Mukesh
While preparing for data science interviews, you
should take notes and also prepare from questions
that have already been asked in companies in
interviews.

For your reference, I have already listed the interview

questions towards the end of this ebook. These
questions have been given topic-wise along with the
name of companies in which these questions have
been asked.

Before mentioning the topics

that are commonly asked in data
science interviews, I would like to
list some key points you must
take care of while preparing for
interviews.

The first question that an interviewer is going to

ask you will be, "Tell me about yourself." Prepare
this question really well. Later in this guide, I'll
explain how to answer this question effectively.
You should be well prepared with your resume.
The interviewer doesn't know about you, what he
has with you is your resume. So, be well prepared
with your resume.
Nitin Mukesh
Now let's quickly reveal all those topics that you
should be preparing for a data science interview.

If you are from Statistics, Computer Science, operation

research, or data science-oriented degrees, chances
are high that you might already be knowing most of
the topics.

But, as I always say, don't forget to revise these. Most

of the time we don't remember what we have studied
a long back. So, without any hesitation, go back to
revise all those concepts from your course notes.

I have also listed some of the books and courses from

where you can read these topics.

Here are commonly asked topics:

Basic Statistic
Probability
Bayes Theorem
Mean, Median, Mode and its properties
Random variables
Skewness, Kurtosis
Expectations and Variance with properties
Correlation and Covariance
Nitin Mukesh
 Statistical Distributions

Binomial distribution
Poisson Distribution
Geometric Distribution
Exponential Distributions
Normal Distributions
Uniform Distributions
Gamma and Beta Distributions
Hypothesis Testing
t-test
f-test
chi-square test
ANOVA
Interpretation of p-value
Central Limit Theorem

Recommended book:
Fundamental of Mathematical statistics
by Gupta and Kapoor
An Introduction to Probability and
statistics by Rohatgi and Saleh Nitin Mukesh
 Machine learning Algorithms
Linear regression
Logistic regression
KNN
SVM
Decision Tree
LDA and QDA
Random Forest
Bagging
Boosting
PCA
Naive Bayes classifier
Ridge Regression
Lasso Regression
Elastic Net
K-Means clustering

Recommended book: ISLR (Introduction to

Statistical Learning)

Nitin Mukesh
 Data Science Concepts
Bias-variance tradeoff
Precision and recall
ROC-AUC curve
Handling missing data
Outlier handling
Cross-validation
Overfitting and Underfitting
Regularization

Recommended book: ISLR (Introduction to

Statistical Learning)

SQL Concepts
Joins in SQL
Constraints in SQL
Primary Key and Foreign Key
Queries in SQL
TRUNCATE, DELETE and DROP
Use of Aggregate Functions
WHERE, HAVING and GROUP BY clause
Nitin Mukesh
 Data Structure Concepts
Array
Linked List
Stack
Queue
Binary search tree
Binary Search
Sorting algorithms

Programming Concepts

List, tuple, dictionary, and set

Loop and function
Recursion
Dataframe
Importing data
Data Visualization Libraries
ML Libraries

Other concepts

Puzzles
Case study and Guestimates

Nitin Mukesh
If you are appearing for a data science profile, most
of the time, questions are based on the topics that
have been mentioned above.

But, as I have told in the beginning that the data

science profile nowadays has been diversified.

The company now hires for more specific roles in

the data science domain, like Machine learning
Engineer, Research Scientists, Time series analysts,
banking companies like ANZ, Citi, Wells Fargo, IDFC
First bank, etc hires for Modelling risk analysts or
financial analysts.

Other product-based companies hire NLP

engineers, Computer vision engineers, Image
processing specialists, etc, which are deep learning
specified roles.

But, deep learning is an entirely different domain.

This ebook is only for data science specialized roles.

In data science specified roles, the interview will be

around these topics+your resume+ Job profile.

Nitin Mukesh
 Make a nice resume: A resume is very important
for any job interview. Resume-making can be a
tedious task, but a good resume is not made in a
single day.

You will have to sit for many days, making small

changes, consulting with your senior or some experts,
and ask for advice.

Make sure that your resume doesn't look

overqualified or underqualified.

You may be wondering, why an overqualified resume

is bad?

Well, being “overqualified” is seen as a bad thing, not

because someone “overqualified” will perform too
well, but because they are expected to be dissatisfied
with the job.

Granted, a person who is “overqualified” will generally

learn their job in a minimal time.

If you are in an overqualified zone, you might be over-

questioned by the interviewer and can have a bad
interview.

I have been rejected from a lot of interviews due to

being over questioned. Nitin Mukesh
So, don't make this mistake. Spend some time making
a good resume. Take help from others. It takes some
time to make an awesome eye-catching resume.

Improve your communication skills: You may be

knowing a lot of things in data science, have done
awesome projects, and are well-versed with
everything, but you are not able to explain
yourself, guess what? You will be rejected.

It is a harsh truth that the company prefers one with

good communication skills rather than one with
complete knowledge.

And if you have both, then you are gold.

But, what if you are not able to explain yourself in

the interview?

This can be fixed. You need to learn to explain things

better.

Record yourself over the phone and watch it, listen to

your recorded voice, see how it sounds. Have
questions discussion with your friends. Discuss in the
group. Have a group mock interview with your friends
in which one will explain the answers to a group.
Nitin Mukesh
With practice, you can easily master this skill. But,
don't ignore it. What students commonly think is, I am
an expert in data science, nothing will happen if my
communication skills are bad.

But, guess what happens?

The interviewer will give preference to those who can

explain. So, start to prepare early, work on your weak
points, and do outstanding in any interview.

Engage yourself in the interview: Don't just sit

answering what the interviewer asks you. The most
important thing is you have to be present in the
interview.

When told, ask questions from the interviewer. Ask

how their journey has been till now. When you don't
know the answer, ask if they can give some additional
hints and then try to solve. This will certainly have a
positive impact on the interviewer. Nitin Mukesh
 Don't interrupt the interviewer: If the
interviewer asks anything and you know the
answer very well, then don't interrupt in between.
Let the interviewer complete the question.

After the interviewer finishes, you can start. I had this

bad habit of interrupting. But, I came to know about
this mistake during one of the mock interviews.

Also, if you don't know the answer to any of the

questions, then you can politely say that "Sir, I don't
know the answer." or "I have not come across this
term till now."

You are not supposed to know everything. I myself

was only able to answer 85% of questions during the
interview for the company I was placed in.

My one of the batchmate who was placed at highest

package from our department was able to answer 80%
of the questions.

The key is that, whatever you explain, be clear, be

specific, use examples and tell an amazing story out of
these.
 Start simple, be clear: When asked any questions,
explain in simple and clear language. Don't
complicate things.

Don't explain things in hurriedness, explain slowly. If

you need time to think, tell the interviewer, "Sir, can I
take one minute to think?"

Simplicity is key.

Be stress-free: If you will be stress-free during the

entire interview, you will be able to think clearly
and answer better.

I still remember my first face-to-face interview for

Accenture. It was for a data science internship. Though
I was very much worried and anxious. But, as the
interview proceeds, I become more comfortable.

But, due to COVID-19, the entire placement procedure

was online. There was a problem with me. I could not
think focus and recall what I had learned in an online
interview. I usually used to get distracted and forget
things.

I had found this thing before the placement procedure

was about to begin. Nitin Mukesh
I used to practice a lot with my laptop camera on. Also,
I used to have online mock interviews with my friends.
These things helped me a lot, and I started to feel
comfortable.

But as placement proceeded, failing a lot of interviews

really taught me how to handle pressure during
interviews. I

It's all part of learning.

Analyze yourself: During preparations, you

should analyze yourself at each step, like what you
need to study, find your weak spots, work on
them.

I advise maintaining a record of what you have to

read, what is still left, and with time update it. Keeping
a record of what you studied really makes things
easier.

After the interview also, write what was asked, what

were the questions that you weren't able to answer,
search on the internet, and read them.

This way, even if you are not selected, now you have
some experience. That's how ML models also learn.
With more data, the models get better. Nitin Mukesh
Do your own research: If you are selected for an
interview, read about the company. Do your own
research. Find about what the company does, how are
their work environment, how is the working culture.

You will find some questions about the company that

HR may ask in HR Round.

Also, if you appear for any start-ups, do find about

them. Sometimes some fake companies do visits and
later revoke your job offer.

At IIT Bombay, a company that was offering 75 lakhs

CTC with 20 LPA gross package on day 14. Everyone
was shocked that any company is offering this much
on the second last day.

But, soon my friends informed me that this company

seems to be fake. We did much online research and
found that this company is really fake and had nothing
to offer.

We immediately informed the Placement cell, and

then the company was questioned. It was ultimately
black-listed.
You know the funny part is the company has only one
employee who was HR, Founder, Interviewer, and
everything. Nitin Mukesh
I know it may sound shocking that such companies
dare to visit IIT Bombay for Placement. But, in the
COVID situation, the fake companies do try to take
advantage. So, be aware of that.

Always do your own research.

Help others and take help: When you get placed,

help your friends in placement. The best part of
life is when you start helping each other.

You should really be grateful for your part and help

others.

You will really find yourself happy when your friends

will also be placed.

Also, don't hesitate to take help

from your friends if you need it.
I was stupid that I took it so
late.. But believe me, put your
ego aside and take help. Give
mock interviews to them.

Life is really awesome. Ego and

jealousy don't have a place in
our life—enjoy the placement
journey.
Nitin Mukesh
 STATISTICS
INTERVIEW QUESTIONS

ASKED IN DATA
SCIENCE
INTERVIEWS
Commonly Asked Topics

Probability Theory
Sampling Theory
Hypothesis Testing
Distribution Theory
Time series Analysis
Design of Experiment

Nitin Mukesh
Prove E(x+y)=E(x) + E(y) (Delhivery)
What strong and weak stationarity? (ICICI
Securities)
Explain ARIMA forecasting step by step. (ICICI
Securities)
What is CLT, z-test, t-test, F test? (ICICI Securities)
When is a t-test used instead of a z-test? (ICICI
Securities)
What do you mean by Replication and
Randomization? (IPSOS MMA)
Trick question on the application of chi-square
test. (ICICI Securities)
Test of equality of proportions. (ICICI Securities)
Structure of a box plot. (ICICI Lombard)
What does correlation mean? What type of
relationship can be found by this? (ANZ Bank)
What is the importance of stationarity in a time
series? (ANZ Bank)
What is skewness in data and how will you remove
it? (LafargeHolcim)
Role of R squared, r, F test in determining features
in Machine learning.
What is stationary time series and why it is
important? (Cognizant)
Define Randomised Block design? Where do we
use RBD? (Cognizant) Nitin Mukesh
What is stratified random sampling and how it is
different from systematic random sampling?
(Cognizant)
What is the p-value? Explain with the help of an
example related to marketing research. (IPSOS
MMA)
What are the assumptions of ANOVA? (JSS, ANZ)
explain CRD (Completely Randomized Design)?
(IPSOS MMA)
What is AB testing? (ICICI Securities)
Consider two fair dice that have been rolled. What
is the probability of obtaining sum 8?(IPSOS MMA)
The lengths of a professor's classes have a
continuous uniform distribution between 50.0 min
and 52.0 min. If one such class is randomly
selected, find the probability that the class length
is more than 51.7 min. (64squares)
Difference between stratified and cluster sampling.
(JSS)
When we use censoring and what are type-1 and
type-2 censoring in Reliability. (JSS)
ANOVA is generally used when we want to test the
significance of the difference between more than
two independent sample means. Why we don’t use
pairwise t-test in this case? (JSS, Accenture)
Nitin Mukesh
What are type-1 error and p-value in inference?
(JSS, Accenture)
What do you mean by hypothesis testing? (Eli Lilly)
There are 3 ants sitting on three corners of a
triangle. All ants randomly pick a direction and
start moving along the edge of the triangle. What is
the probability that any two ants collide?
(Accenture, Delhivery)
You have 9 red balls and 1 black ball in a bag. You
are picking balls 10 times with replacement. What
is the probability of getting at least one black ball
from the bag? (Accenture)
Explain Central Limit Theorem. (ANZ, Wipro)
What is the normal distribution and what is the
shape of the curve? (WNS Global)
What is the relationship between media, mode and
mean in positively skewed data? (ANZ)
What are statistic and data science? (ANZ)
You have to estimate the average height of the
population of India, how will you go about it? (ICICI
Lombard)
What is the use of binomial distribution and its
PMF? (Wipro)
What do you mean by selection bias? (Wipro)

*** Nitin Mukesh

REGRESSION ANALYSIS
INTERVIEW QUESTIONS

ASKED IN REAL
INTERVIEWS
1.What are the assumptions of linear regression?
(Delhivery, ANZ bank, Citi Bank, Accenture )
2. What is the meaning of multicollinearity? (ANZ,
Amazon)
3. How to detect multicollinearity? (Amazon,
Delhivery)
4. What do you understand by VIF (Variance Inflation
Factor)? (Amazon)
5. What is the difference between R-squared and
adjusted R-squared? (Delhivery, ANZ bank, Citi
Bank, Accenture)
6. How to deal with multicollinearity in data? (Citi,
Accenture)
7. Explain forward and backward elimination?
(Accenture)
8. How PCA works? (ICICI securities, Amazon,
Miko.ai)
9. Explain Ridge and Lasso Regression? (Delhivery,
ANZ bank, Citi Bank, Accenture, Amazon)
10. Can SVM be used for regression? (Miko.ai)
11. What is the curse of dimensionality? Can you give
an example?
12. What is the difference between the coefficient of
determination, and coefficient of correlation?
13. Give methods of variable selection in Regression
Analysis? (Delhivery, ANZ bank, ICICI securities)
14. Why do we perform the residual analysis? (ANZ)
15. What are L1 and L2 penalization? (Miko.ai)
16. What is heteroscadasticity? How does it affect the
regression coefficients? (ANZ)
17. Why does only VIF>10 implies that there is
multicollinearity, why not choose vif>8? (IDFC First
Bank)
18. In my dataset, if I have 100 observations and 1500
features, do you think whether I would be able to fit
the regression model onto that or not? (IDFC First
Bank)
19. For a single variable, how will you detect outliers?
(ICICI Lombard)
20. How correlation between two variables will
change in presence of an outlier? Will it increase,
decrease or remain constant? Explain how, using its
formula. (ICICI Lombard)
21. What are influential and leverage points? Which
of them have more effect on the model? (ANZ, Wells
Fargo)
22. Does multicollinearity impact the prediction of a
machine learning algorithm? (Wells Fargo)

***
 LOGISTIC REGRESSION
INTERVIEW QUESTIONS

ASKED IN
PLACEMENT
 Logistic Regression
Logistic regression is a supervised Machine
learning algorithm used for classification.
It is mainly used to model a binary dependent
variable.

Logistic Regression can be easily used to classify

more than 2 classes.

Example of Binary Classification,

To predict whether an email is a spam or not
Whether the tumor is malignant or not

Nitin Mukesh
What is the difference between linear and logistic
regression? (ANZ. Wells Fargo, Accenture)
Can you use Logistic regression for regression
problems? (Accenture)
Explain the working of Logistic Regression? (ANZ,
Delhivery)
What do you mean by Generalized Linear Model?
(ICICI Lombard)
Is logistic regression a Generalized Linear Model?
How Support Vector Machine different from
Logistic regression?
Write the expression for the Logistic function.
What do you mean by deviance in Logistic
regression? (ANZ)
What are penalized logistic regression models?
Why did you use them? How are they better than
logistic regression models? (FCS Limited)
Differentiate between Lasso, Ridge, and Elastic
Net. (Wells Fargo)
How to handle imbalanced data in classification?
Which metric to use in this case? Why is accuracy
not a good measure in this case? (FCS Limited)
How will you deal with overfitting in the case of
Logistic Regression? (Delhivery)
Interpretation of parameters of the logistic
regression model. (ANZ) Nitin Mukesh
Steps to build a Logistic Regression model from
scratch. (ANZ)
If the accuracy of the model is 95%, is it good?
(ANZ)
Difference between Logistic regression and
decision tree. (Delhivery)
Difference between Logistic regression and
Random Forest. (ANZ)
What do you mean by cost function? (Amazon)
What is a convex function? (Amazon)
Explain Gradient Descent Algorithm. (Meru cab)
How will you use logistic regression if your data
has more than 2 classes?

***

Nitin Mukesh
INTERVIEW QUESTIONS
ON CLASSIFICATION
ALGORITHMS
 Commonly Asked
Topics

Logistic regression
KNN
SVM
Decision Tree
Random Forest
Bagging
LDA & QDA
Naive Bayes classifier
Classification Metrics
Overfitting and underfitting

Nitin Mukesh
Difference between logistic and SVM. (Accenture)
How to deal with overfitting in Logistic Regression?
(Delhivery)
If the accuracy of the model is 95%, is it good?
(ANZ)
Does standardization have any impact on the
performance of the model in the case of a random
forest? (Wells Fargo)
How to handle imbalanced data in classification?
Which metric to use in this case? Why is accuracy
not a good measure in this case?(FCS Limited,
Citi)
Explain the Logistic regression model. (ANZ,
Miko.ai, IPSOS)
Suppose you have 200 variables, will you consider
all the variables to build a model? (ANZ)
Difference between random forest and logistic
regression? (ANZ)
What is a random forest and how it differs from
Adaboost? (Meru cabs)
Explain bagging in Machine learning? (TCS
innovation lab)
Explain F score, precision, recall, ROC curve. (Meru
cabs)
When can you say, KNN model is overfitted?
(Delhivery)
Write the loss function of logistic and linear
regression. (Delhivery)
How is the prediction made in the random forest?
(FCS Limited)
How is XGBoost different from gradient boosting?
(Wells Fargo)
What will happen if we use MSE in logistic
regression? (Delhivery)
If you have categorical features in your data, then
how will you find the Nearest Neighbor using KNN?
(ANZ)
Explain the working of the Support Vector Machine
Algorithm? (ICICI Lombard, Cognizant)
What is the kernel trick in SVM? Why is it
computationally efficient? (FCS Limited)
Explain the complete working of the decision tree
algorithm. (WNS Global, TCS innovation lab)
What is a confusion matrix? (ANZ)
Difference between Linear regression vs logistic
regression. (Wells Fargo)
What is underfitting and overfitting for example?
(Accenture)
What are the steps to build a model from scratch?
(ANZ)
Why random forests are called random? (FCS
Limited) Nitin Mukesh
Can Ridge or Lasso be used for NLP? (Miko.ai,
NLP)
What is the difference between Bagging and
Boosting? (Miko.ai, TCS Innovation Lab)
Differentiate between SVM and Random Forest?
(Miko.ai)
Which is better, Random forest or decision tree?
(TCS Innovation Lab)
What is the Ginni index and entropy in the decision
tree? (Cognizant)
What do you mean by deviance in Logistic
regression? (ANZ)
What is a convex function? (Amazon)
What are the criteria to evaluate a Logistic
regression model? (MMA IPSOS)

***

Nitin Mukesh
Commonly Asked
Topics

Loops
Functions
Recursion
Dynamic Programming
Sorting algorithm
Binary Search
Coding questions are commonly asked in data science
interviews. But, if you are sitting for banking
companies like ANZ, Citi, ICICI, IDFC first bank, etc, you
can expect no or minimal coding questions.

Companies like Delhivery, Accenture, TCS, Flipkart,

Microsoft, etc ask a little bit more in-depth questions
from coding sections.

But, if you are comfortable with the topics, I have

mentioned above, then you are good to go. More is
better, but at least prepare that many of things.

I bet you can easily get these topics from YouTube. If

you are interested in reading these topics in-depth,
the channel that I recommend is given below:

Codebasics: Link

Don't worry if you are not an expert in coding. As I told

you, something is better than nothing. At least read
the basics if you can't master all these topics.

Many times, questions based on the application of the

basics of loops are asked. So, at least prepare the
basics.

Here are some questions that have been asked in the

data science and machine learning engineer interview.
Write a code to find unique integers and the count
of an array. (Ignitarium Solutions)
Explain unsigned and signed integer
representation in binary. (Ignitarium Solutions)
You have a list of natural numbers. Using that,
make a list of even numbers in just 1 line of code
using python. (FCS Limited)
How to make a function with variable number of
parameters in python? (FCS Limited)
How do you merge data frames in Pandas? (FCS
Limited)
What is the use of groupby function in Python?
(FCS Limited)
You have a list of names of cities. A city can occur
more than once on the list. Write a Python code to
count the number of each city. (Accenture)
You are given two sorted lists you have to merge
them in a single list in sorted order in O(n).
(Accenture)
There are n steps in a stairs. You can hop either
one or two steps at a time. Write a code to find the
number of ways one can reach the nth step, given
n. (Accenture)
What is pandas used for? (TCS Innovation lab)
What is a data frame? (TCS Innovation lab)
How will you create an empty data frame in
Python? (TCS Innovation labs)
What are the different types of loops in
programming? (TCS Innovation labs)
How will you detect an infinite loop in a while loop?
(TCS Innovation labs)
Write code for Binary Search. (Delhivery, Cargill)
How do we delete an element from a linked list?
(TCS Innovation labs)
Write a code to find the mean and standard
deviation of a given list of numbers, without using
libraries. (AIRA Matrix)
How did you import a module from the library in
python? (LafargeHolCim)
What is the difference between an array and a
linked list? (TCS Innovation labs)
Write a recursive and non-recursive code for
preorder tree traversal. (Cargill Data Engineer)

***
 ASKED IN
PLACEMENT
 Why Guestimates?
We’ll start by discussing guesstimates, for which
candidates are asked to come up with a figure,
usually the size of a market or the number of
objects in an area or to estimate a particular data.

Although guesstimates are more commonly given

in interviews for consulting positions, they do pop
up in finance and analytics interviews as well.

Practicing guesstimates is a good way to begin

preparing for stress questions in finance
interviews, as they force candidates to think
aloud – precisely what interviewers want to see.

Nitin Mukesh
 How to solve?
The most important thing to remember about
brainteasers, guesstimates, or even simple math
questions that are designed to be stressful is to
let your interviewer see how your mind works.
And, that's what your interviewer wants to see.

The best approach for a guesstimate question is

to think of a funnel. You begin by thinking
broadly, then slowly narrowing down the
situation towards the answer. Show how you
reach to answer.

For any guesstimate or brainteaser question, you

will need to understand whether your interviewer
will be providing any direction or whether you will
have to make assumptions.
Nitin Mukesh
 An Illustration

e.g. Let us say you

have to estimate the
number of trees on
your college campus.

You can start by asking few initial questions to the

interviewer like, will I have to estimate full-grown
trees or small trees too?
This shows you are trying to understand the
situation.

Then proceed with dividing your campus area into

two parts, heavily populated with trees and less
populated with trees.
Then, you can make an initial estimate that let us
say per 100 square meters in densely populated
areas, 40 trees are there and hence calculate total
trees by dividing the area of the densely populated
region. The same goes for the other one.

Nitin Mukesh
 Conclusion
See, you don't have to be exact in your answer.
But, you have to show how can you break down a
problem to solve it efficiently. That is what
interviewer is interested in.

With this, let's see some of the guestimate

questions asked in real finance and analytics
interviewers by different comapnies.

Nitin Mukesh
Amount of revenue from the mobile industry in India
(KPMG)
Estimate the number of trees on the IIT Bombay campus.
Number of Maggi consumed on campus in a month.
(Deloitte)
Number of smokers in Hyderabad (Deloitte)
Number of Vada Paos served in the XYZ Mess (Deloitte)
Number of Maggi packets sold on the first day of the
relaunch (Deloitte)
Number of flights taking off in a day from the Delhi Airport
(Deloitte)
The number of taxis in Bombay. (Deloitte)
Number of people who have ever lived on the Earth
(Deloitte)
Number of vehicles in a toll gate of your choice (Amazon)
Number of calls to our customer care in a week (Amazon)
Number of people using FB on campus (Amazon)
The market for cricket bats in the country (Amazon)
The market for leather shoes (Amazon)
Estimate the number of people on Patna Gandhi Maidan.
(ANZ)
Market size for electric insect repellant (KPMG)
Estimate market size of electric bulbs in India (KPMG)
Guesstimate number of daily flights in India (KPMG)
Estimate market size of online retail in India (Amazon)
Estimate the number of office chairs in Delhi (EY)
Estimate number of autos in India (Accenture)
Estimate number of Gmail app users in India
(Accenture)
Nitin Mukesh
Number of ticket counters required for Lucknow metro
station (Amazon)
Titan is launching a watch priced at Rs. 12,000. Estimate
market size (Accenture)
Royal Enfield is launching a new bike priced at 2 lakh in India.
Estimate market size. (Deloitte)
Estimate Revenue of Make my Trip in a year (Deloitte)
Number of people at Lucknow airport in a day (i3)
Number of cabs required at Lucknow airport (i3)
Calculate the total amount of revenue from the mobile
industry in India (KPMG)
Estimate the amount of food wasted inside IIT Bombay
hostels. (Accenture)
You have to estimate the average height of the population of
India, how will you go about it? (ICICI Lombard)
Suppose the vaccine for Corona is ready. How will you plan
its delivery? (ICICI Lombard)
Estimate the number of masks which are disposed of every
day in a city. (ANZ Bank)
Estimate the number of flights going from Mumbai airport in
a week? (ANZ Bank)
Estimate the number of persons you need to run the
campaign at the airport? (ANZ Bank)

***

Nitin Mukesh
Puzzles for Data
science interview

ASKED IN
PLACEMENT
 How to answer puzzle
questions?
When the interviewer asks you a puzzle, you don't
need to tell an answer directly or come to conclusion
at once.

Explain how you arrive at the answer. Explain the

step-by-step procedure of getting to the answers.
There are a lot of puzzles that are commonly asked.

I'll list some sources from where you can read

puzzles.

Try to read 2-3 puzzles daily. Don't see the answers

directly, just try to think for some time and then see
the answer and analyze if you are correct.

I used to read 3-4 puzzles daily before going to sleep.

This way I was able to cover most of the puzzles in 2
months.

Here are some sources from where you prepare

puzzles:
Logically Yours (YouTube Channel): Link
GeeksForGeeks Puzzles: Link
Here are few puzzles that have been asked in data
science interviews:

The lengths of a professor’s classes have a

continuous uniform distribution between 50.0
min and 52.0 min. If one such class is randomly
selected, find the probability that the class length
is more than 51.7 min. (64 squares)

There are 3 ants sitting on three corners of a

triangle. All ants randomly pick a direction and
start moving along the edge of the triangle. What
is the probability that any two ants collide?
(Accenture, Delhivery)

You have 9 red balls and 1 black ball in a bag. You

are picking balls 10 times with replacement. What
is the probability of getting at least one black ball
from the bag? (Accenture)

Suppose that we wish to know which stories in a

100-story building are safe to drop eggs from, and
which will cause the eggs to break on landing.
What strategy should be used to drop eggs such
that the total number of drops in the worst case is
minimized and we find the required floor?
(Delhivery)
You are provided with twenty-five different
horses and you must find out who are the fastest
horses. You can conduct a race of five horses only
at one time. There is no point in the race where
you can find out the actual speed of a horse in a
race. How many races will it take to help you
determine the fastest three horses? (ANZ)
There is an ant on the top corner of a room and it
has to travel to the diagonally opposite bottom of
the room? What's the shortest way possible?
(ANZ)
Given two hourglass of 4 minutes and 7 minutes,
the task is to measure 9 minutes. (ICICI
Lombard)

***
HR QUESTIONS FOR
DATA SCIENCE
INTERVIEW
ASKED IN
PLACEMENT
 Why HR Questions?
After the initial rounds of candidates selection, every
company holds an HR interview round to get to know
the potential candidates personally.

An HR interview’s main aim is to estimate the

candidates’ personality, background, strengths, and
weaknesses and assess whether or not one is a good
fit for a specific role.

In other words, the HR round decides if you’ll be

called in for the following rounds.

Nitin Mukesh
Though there can be a variety of questions that can
be asked in the HR round depending on the job role
and your degree. But, here are some questions that
are more common in any interview:

Tell me about yourself.

Why should we hire you?
How do you see yourself in the next 5 years?
What are your strengths and weaknesses?
Why do you want to join this company?

These 5 questions are the most common interview

questions asked in the HR round. But that's not all.

To give you a glimpse of what kind of questions can

be asked in the HR round, I have listed some
questions below for your reference. These questions
have been asked in different companies like
Accenture, EY, ANZ, TCS, Citi bank, Wells Fargo, ICICI,
UHG, etc.

You don't need to write down answers to all these

questions. Just discuss some of these questions with
your friends during preparation.

The key is to be well prepared.

Tell me something about yourself.
How's your technical interview, rate yourself out of
10?
Why statistics/data science?
Have you ever used data science in your real life or
any day-to-day application of data science?
What is the biggest problem in India, how can you
contribute to it as a data analyst?
What are your expectations from our company?
What are your long-term plans?
What are your strengths and weakness? (at least 3)
How are you doing to overcome your weaknesses?
What are your hobbies?
Where do you want to see yourself after 5 years in
our company?
According to you, a big reputed company matter
or high salary in a start-up?
What do you prefer - Teamwork or individual work
What are the ethics, an employee should possess?
Do you have managerial skills?
Do you any issues regarding job location?
What is your greatest fear?
Do you have any questions from us?
You are from statistics, why do you want to join IT
Sector?
What is your main motivation behind this profile?
Any failure in life and what did you learn from it
and how you succeeded afterward?
What do you understand from this job profile?
What's your aim in life?
Where do u see urself in 10 years not from a
carrier perspective but as a personality?
Something u tried very hard to achieve but
couldn't achieve?
2 things that you like about your college?
2 reasons that make you suitable for this profile?
Why do you want to join a bank?
Would u be ready to move abroad if given the
opportunity
Why are you interested in joining the finance
sector?
How will you deal with the problem of no
background in finance?
What makes you wake up in the morning?
Suppose your co-worker working on a project with
you drops the project in between. How would you
make sure that the project is completed before its
deadline?
Nitin Mukesh
Why choose the corporate sector after having a
good academic report?
What are your thoughts on corporate life?
How do your friends define you?
If you have infinite time and money what you'll do?
Have you come across a situation when you have a
proposal and your council members are not ready
to listen to you?
How do you manage events?
What steps you have taken when everything in
your life is not going well?
Suppose one company is offering you the work
you like to do and with a decent salary and
another company gives you a higher salary but you
don't like the work in that company then which
company you will choose?
Tell me about how you feel when you think in the
past that if you would have improved certain kinds
of things then the result could have been better
and give an example.

***

Nitin Mukesh
It is the most important question in the interview. This is
the first question you will be asked. Though, if asked in
the technical round, the answer will be different than
what you say if asked in HR Round.

I still remember my first interview for an internship.

When the interviewer asked me, tell me about myself, I
was not able to answer. But, somehow I managed to tell
about my educational background and all. Then I was
blank.

Later I came to know that all the students prepared for

this question prior to the interview. I was stupid there
who had no idea about it. So, prepare this question really
well.

Here are the main components of the answer:

Educational Background
Projects and internships
Your strength
Conclusion

Let us see how can you make an ideal answer to this

question.

So, here is a sample answer to this question:

 Educational
Background

"Thanks for giving me this opportunity to introduce myself. I

have done my graduation from BHU in computer science and
my MSc from IIT Bombay in Applied Statistics. (You may tell
about your school education too) Projects and
Internships

If I talk about my professional experience and projects, I have

done a machine learning research internship at IIIT Allahabad
on self-taught learning using autoencoder.

I have done three projects, My MSc project was on Statistical

analysis of proteins of COVID-19 patients, My other project was
on the prediction of cancer mortality rate for US countries
based on the application of Regression analysis.

My third project is on-time series and statistical analysis of the

covid-19 scenario in Pune. (You can talk about
competitions you have won or Position of Responsibility
Your
here) strength

I'm proficient in Data Science, Machine learning, and coding

concepts using Python. I am good at picking new technologies
fast and I think time management is my strength. Conclusion

And I feel I am now ready for a more challenging assignment

and the prospect of working with XYZ firm really excites me a
lot."
Well, this is the answer that I told during my interview at TCS.
"Awesome introduction, Nitin." the Interviewer replied.

Prepare your answer well, record it and listen to it again

and again so that it should sound perfect. Make sure
that it should be between 60 seconds to 90 seconds.

The answer to the same questions will be different for

the HR Round.

Here are the main components of this answer:

Educational Background
Strength
Hobbies
Interests
Conclusion

Education background and conclusion will be the same.

In the strength part, you have to tell about your strength
like time management, fast learners, dedication, etc, then
your hobbies and interests like music, sports, traveling,
etc.

Yes, the ideal answers should lie between 60-75

seconds.

Prepare two types of answer for this question in advance

and nail the interview.
You prepare for the campus placement, but not
everyone gets selected in the very first company.

But, you don't need to get disheartened 😔 . The

campus interview sometimes can be full of ups and
downs. You don't know which is your lucky day.

Despite being prepared, I was not able to clear the

interview during my campus placement season. I was
able to clear most of the coding and aptitude tests,
but in the interview round, I was getting rejected.

Sometimes the interviewer

praised me for the interview
round, but the company didn't
respond.

During those hard times, it feels like everything is such

a waste. What you have been doing till now is of no
use.

You see your peers getting placed and here still you
are waiting for another chance. Maybe someone with
no experience or less technical skills than you is
getting placed and still, you are not.
Self-doubt starts to kick in and you will start thinking
that maybe you are not made for this job. All those
lucky people got selected and you didn't.

But believe me, there is a lucky day for you. You can go
through all sadness. It's the consequences of what is
happening to us.

You need to keep going. Prepare even more. Analyze

your mistakes. Improve them. Sometimes, there may
not be your fault, accept it.

The road to success is not easy and not fair for

everyone. After each interview, write a short note on
how you feel about this interview? Think if you need to
prepare more. Write those questions that you couldn't
answer. Revise those topics that you were not able to
explain clearly.

Improve yourself. You have to potential to get through

it. Everyone has to go through this pain sometimes.

I had signed 110 companies, selected for tests in 40+

companies, interviewed 17 times before I got placed
as an ML Researcher at TCS. And there is still
someone who got placed even later.
I remember a story my brother was telling me when I
was continuously getting rejected from interviews.

"There was a student in his college. His coding and

aptitude skills were good. He had cleared tests and
appeared for 100+ companies during college
placement before he got placed. "

Keep going. You are going to learn a lot when one day
you will be placed.

I'm able to write these things because I have

experienced the mindset of a student who is getting
rejected in every interview.

The key to an interview is simple: Prepare well, present

yourself even better.

Put yourself 100% in the interview. Be confident and

explain clearly.

That's all.

All the best for your next data science job interview!!!
😇

***
 For more Data science placement
preparation resources, Please scan:

I would like to ask you for a moment to

fill this feedback form:

https://forms.gle/JNMLr6zK6yYAo7UG9