JP Morgan and Chase Interview Experience

I was shortlisted for interviews of both quant and Corporate Investment Banking profiles at JP Morgan and Chase through a campus internship.

1 CIB and market research:

There were 3 rounds in total.

Round 1:

The interview started with a small introduction where I told about my branch(non-cs branch), my dual degree specialization, and institute co-curricular activities in the coding field. Then he asked me if I had any idea about finance, I told him that I read some blogs but I don’t have any relevant experience or courses in finance. Then he asked me the following questions:

• Write a code for any sorting algorithm on a paper: I wrote quick sort.
• What is the difference between R-squared and Adjusted R- squared: I was able to answer it as I did a statistics course.
• Define P value and what in null hypotheses: I was able to answer this question.
• He asked me a question on poisons distribution, it was something along the line of ‘if a book has a total of these many errors, what’s the probability that the particular page and given a number of errors, as soon as I mention that the problem can be solved using Poisson distribution, he said my answer is correct and no need to explain further.
• Then he started asking ML questions, as I mentioned ML many times in my resume
• I had a credit fraud analysis project in my resume so he asked what were the parameters for the data
• what problems did u face with the data set. I told them the data was kind of biased as the number of fraudulent transactions was very less compared to correct ones, so I had to over-sample the data, then I explained how I did it using SMOTE
• I also told him I removed unnecessary columns from the dataset after calculating VIF(Variance inflation factor) after which he asked me to write the formula and explain what VIF Is, I wrote the formula and said if is used to detect multi-collinearity, as soon as I mentioned it he said right answer and mover to the next question.
• After this he asked a few finance questions: 1) What are bonds, 2) Do you know what are options, explain a call and put options+a case study on call and put options. I was able to answer these as I knew the terms
• He asked me for the latest financial news, so I told him what happened at Silicon Valley Bank, he asked me how it was different from the 2008 crisis, So I told him the difference as I saw a few videos on both the topics before interviews.
• He seemed pretty happy with the interview so he sent me to the next round.

Round 2:

• This round was more or less resume based and the interviewer was really chill, he pretty much didn’t ask me any specific questions, just a bit about my resume point.
• Which ML also gave the best accuracy for your project?
• Explain SVMs
• He asked about my research project at TUM Germany.
• Tell me about DBSCAN and why not KNNs.
• Write a sorting algorithm

Then he told me that they have selected me for the CIB role, but as I have filled the quant role also sent me to the quant interviews.

2) Quant research:

Round 1

• Tell me about yourself
• What’s your favorite data structure, I said it was Binary Trees, the interviewer didn’t seem to like trees so he skipped to the next question
• He asked me to explain linear regression, I gave him an explanation.
• He told me to write the error function for regression and explain whats the intuitive meaning of w in linear regression. I told them it meant the weight of all the features respectively, he was satisfied so he moved to the next question.
• Tell me about logistic regression. hows w in logistic regression is different from w in linear regression, I couldn’t understand the question at first, but after he give a few hints I was able to answer correctly so he seemed satisfied.
• Why quant?

Round 2

• Tell me about yourself.
• Explain SVMs, soft margin SVMs
• Does linear regression have a closed-form solution?
• What is gradient descent, and why do we use gradient descent in linear regression?
• What is logistic regression, and what is the gradient of logistic regression?
• Does logistic regression have a closed-form solution? is gradient descent necessary for logistic regression?
• Tell me about the DBSCAN algorithm.
• He asked me if I knew SQL, I said no so he moved on.
• He asked a few questions about why quant.
• This was pretty much the quant round, then they sent me to the common hr round.

HR round: What profile do you prefer CIB or Quant? Do you have any other offers?

They told me I am selected for both roles and I can choose anyone, I chose quant as my resume was more coding-oriented and quant seemed more interesting.

I got an offer for the QR role.