Hi,I have been invited to the first round of interviews for a PhD Quant Research position. Does anyone have any tips for the telephone interview? I am a CS PhD and have not come across much financial mathematics and I certainly need to refresh my mind on the likes of SDE PDE and probabilities. Should I spend more time on preparing for the C++ and algorithmic questions or on maths and finance stuff? My research is in applied machine learning, although this was not mentioned in the job brief, will they ask me ML and Data-mining Q's in a telephone interview? Recently I had an interview which I was asked on assmuptions and types of Hidden Markov Models which I have come across several times but did not know of the top of my head. Has anyone been asked technical and detailed ML Q's like this for quant jobs?Thank you

Hi,I don't know what they will ask you but if you have put something in the CV the guy in front will expect you to do talk about it.Ps: I remember 2 years ago, I had to prove liouville theorem in a face to face interview just because I have putted complex analysis in my cv and the guy in front was a Phd in the pure subject

I have been asked pretty detailed stuff before. Part of the point is that they want to know where your knowledge 'bottoms out'. Its hard to say without knowing more about the precise role, but here are some pointers I have found useful:- If its on your CV, it is fair game and if the person asking the questions knows anything about it, then it could get detailed. Hence you must know the topics on your CV in depth. - If its not on your CV, but is in the same general area then you will find that they will ask you fairly simple things. Normally the sort of things that you get in the first section or two of text book chapters on the subject. So the question about the definition of a Markov process is a good example: its a core piece of information. So for revision, go over the important, canonical foundation knowledge. For example in machine learning: know Gaussians inside out, know rules of probability and be able to use Bayes reflexively for simple inverse probability problems. On the other hand you are very unlikely to get asked complex stuff about Variational Bayes and Chinese Stick breaking processes etc... (unless its on your CV). If you proclaim to have expert knowledge in kernel methods but are not fluent with regression then it will look bad. So drill basics. Watching tutorial videos on you tube and doing a couple of questions can be a good refresher on these things. - Dont learn anything new. Stick to what you know. Its ok to say that you dont know something if you have no clue (as long as it is not something you should know given your CV). Get fluent in what you want to present as your competencies. Dont go crazy and think you need to learn everything in stats in 48 hours. You wont be able to.- They might ask brainteasers. They can be embarrassingly easy or uncomfortably difficult. Practicing them can help to warm your brain up. But remember to be cool, think through it, tell them what you are thinking and do your best. Part of this is to see how you respond to the pressure. So I would consider the following to prepare: Specific prep: Go through your CV. If you were an interviewer what would you ask? What projects/keywords would you pick up on? How will you answer these questions? Almost always you will be asked to talk about your research. You must have a clear, concise but informative and interesting explanation that you can tailor a bit to the level of expertise of the person you are talking to (if you dont know it already you should find out their background, if you cant find out before hand, then ask them in a polite conversational way). Focus this explanation around the features of your project(s) that are relevant to that role.General prep: I would decide what you need to know based on your CV and your general areas of competency. Choose good text books for those areas (not hundreds, but a good, easily readable, useful text for each competency) read the intro sections of the relevant chapters. Do some of the questions. Watch some videos about the material on you tube (where possible and relevant). Practice some brainteasers ('Heard on Wall Street' by Crack is a good start, but there is also the brainteaser forum on here and many other sources such as quant net, Mark Joshi's book etc...).Its easy to focus on the technical stuff (you are a quant after all right?) but remember you also need them to want to work with you (potentially). So you need to be amiable and easy to talk to. Dont be surprised if they also ask you some 'soft' questions for this reason. Good luck PS: The finance/math/programming balance depends on the role and your background. If you have said that you know C++ and they have said they want it, then you will get C++ questions. Mainly these questions concern advanced topics in object oriented programming. Algorithms questions are also very popular, for a quick spin through the most pertinent stuff check out: topcoder. Generally knowing about finance is a bonus but is a minor issue compared to the other things. But you need to know enough that you can tell them about the company about the role and about why you want to do it.

I think at the very least you should know what job you are going for, and therefore what is relevant to them.eg is it for derivatives pricing - where ML is not used or for stat arb where ML is ( and SDEs are irrelevant)

QuoteOriginally posted by: spv205I think at the very least you should know what job you are going for, and therefore what is relevant to them.eg is it for derivatives pricing - where ML is not used or for stat arb where ML is ( and SDEs are irrelevant)It'd be good to know some general brownian motion and stochastic process stuff tho.