I applied online. The process took 4 weeks. I interviewed at J.P. Morgan (Londres, Inglaterra) in Jan 2015
Interview
3 stages, 2 Phone interviews and one on-site interview. First 2 interview are over the Phone las about an hour and a half each. Its a very technical interview solely quantitative. One is expected to read John Hull, 8th edition
Interview questions [1]
Question 1
Phone interviews concentrate quant question, mathematics, programming. Very technical interview.
Eg. What is Brownian Motion? Characteristics of Brownian Motion, Derivatives Modelling and Pricing related question.
Lame interviewer who was vdry belliting and showed no interest in finding out about my aptitudes, and very confrontational at times. The interrier didn't last too long and he was very inept and un knowledgeable of the subject matter he was asking about.
Interview questions [1]
Question 1
What's an abstract class, what's polymorphism, some questions regarding my background in quantum mechanics but he obviously could not comprehend a thing I said. A waste of time more like.
I applied through college or university. I interviewed at J.P. Morgan
Interview
Got my resume forwarded by a friend of mine JPMorgan. Immediately my phone interview was setup
Interview questions [1]
Question 1
All the questions were very basic, not the ones your would expected for a PhD position.
1. Number 1-10, of which 3 numbers are chosen without replacement. what is the probability that the sum of these numbers is divisible by 3.
2. Three identical metal balls, when tossed in air come back in time T. After first ball, the second is released when the first ball is at its peak and then third ball is released when the first and second collide. After how much time the first ball is back ? All collisions are assumed elastic. (ans: 3/4T+T)
3. How to you estimate a derivative from discrete nodal values on a 1D nodal mesh starting from 0, h, 2h,...etc where h is the interval between the nodes and the derivative needs to be second order accurate at all nodal points. (Ans: Easy except for initial node '0', for node '0' use taylor series expansion)