Company – IDFC First Bank
Position – Full-time Job
Profile – Analyst BI
Process – IDFC conducted an aptitude test (with verbal and LR) followed by one interview (both technical and HR)
Q1) Tell us about yourself?
Q2) You’ve cleared many actuarial papers, then why do you want to switch your career from Actuarial science to data science? Don’t you think that you’re overqualified? If yes, won’t you leave our company as soon as you get some experience from here?
Q3) Tell us about your Regression project?
Q4) How do you test for the presence of multicollinearity and how do we remove it?
Q5) Why only VIF>10 implies that there is multicollinearity, why not choose VIF > 8?
Q6) What is the formula to calculate VIF? How do you interpret VIF?
Follow up: How do you calculate the R-square which is used in the formula of VIF? (not the normal R square)
Q7) In a dataset if I have 100 observations and 1500 features, do you think I would be able to fit a regression model on it? If yes, will it be fine? If no, why?
> I said yes
Follow up 1: He said okay that’s fine but do you really think it’d be good if we do that?
> I said no
Follow up 2: Why?
1) Interpretability would be totally lost
2) Significance of predictors wouldn’t be that high even for the predictors which are significant in reality because our equation would be Yi = b0 + b1X1 +…. + b1500 X1500 + Ei.
3) Since rank won’t be more than 100 of the feature matrix, we would be getting infinitely many solutions and which machine would give us would be one of those infinitely many solutions.
Q8) What’s the difference between R-square and adjusted R square?
Follow up 1: Why do we use only adjusted R-square? Why don’t we go for some other accuracy metrics like AIC, BIC or Roc curves, etc?
Q9) interpretation of recall score?
Then, he asked HR to ask me some questions and HR said that no sir, I’m also done with him, I don’t wanna question him on anything.
I was like what the hell!! At least ask me one question and instantaneously I felt like this interview is also gone from my hands and it’s over now, I couldn’t make it.
But it was their trick, to take me to this level of tension and then they can throw a bomb at me by asking an amazing case study!
So, the technical interviewer started again and he said that lovish before you leave, I have one last question for you!
Q10) Since we are a bank, we need to decide who are our potential customers so that we can offer loans. To decide that (we didn’t have any data with us), we collaborated with Uber cab services and they agreed to give you data of all customers who took rides from them in the previous 2 years.
Now, tell me, what are the important features you’ll look at for deciding the set of people to whom you can offer loans?
Ans.) I said I’ll offer loans to those who are Financially better because if I offer a loan to someone who is not even earning that much or doesn’t have that good financial base, it would be more likely to be bad debt for our bank.
And from the data, these important features could be:-
– Frequency of rides
– Money Spent on rides
– Ratings given by driver to the customer (which is secret to customers, only uber has that)
– Mode of payment( who pays in cash and who in credit)
And finally, he asked the last question:
Q11) Do you have any questions for us?
I asked what I should improve because if I get a chance to work at IDFC, I’ll still have 6 months to groom myself further according to the needs of the IDFC bank. He suggested that I improve my coding skills.
Then he asked me, how do you rate yourself from 0 to 10 in coding.
This is how the interview ended.
For any query, suggestion or feedback, please reach out to us on LinkedIn. You can find some of the resources that helped us here.
If you can contribute by talking about your interview experience, it will definitely create an impact. Please fill this form to help students get a perspective about the interview structure and questions.
You can read other articles here.
Cheers and Best!