5 Problems with Quant Interviews and How to Correct Them
We’ve all been there. That difficult, make-you-sweat-in-your-seat interview is everyone’s nightmare. Unlike traditional behavioral interviews conducted at most companies, interviews for quantitative finance (quant) jobs are a different beast altogether. Quant interviews are a gauntlet of highly technical questions which you must know the answer to straight away. Total recall, all the time. Here are five problems with how quant interviews are conducted and how to correct them.
1. Stochasticity of Questions
Quantitative finance is the nexus point of several fields: finance, mathematics, economics, statistics, computer science, and engineering. It is no wonder it employs the top intellectual talent on earth. From physicists, to chemists, to mathematicians, to economists, or to electrical engineers, quantitative finance is a highly diverse field with highly diverse talent.
What tends to happen during quant interviews is that interviewers try to test bits and pieces from each of these subfields. This ends up creating two problems. The first is the same as with every other type of interview: the interviewer has his or her favorite questions which they ask each time without variation, which the interviewee does not know ahead of time. Fair enough. However, this spirals into the second problem which is the absolute randomness of topic area.
I once interviewed for a desk quant position on a municipals trading desk. The headhunter told me to be prepared for questions regarding stochastic calculus and Black Scholes model variations. I did so, fully re-studying and recalling what sigma algebras and filtrations were all the way to different variations of stochastic models that could arise. I knew my stuff, inside and out. I come in to the interview only to get hit with a softball first-year graduate econometrics question; something that had been collecting cobwebs in my brain for 4 years at this point. I literally hadn’t thought about this topic in ages. Thus, I ended up struggling through the problem which was a waste of time for both the interviewer and me. Remember, in order to secure a job in this field, you must have total recall at all times during the interview, no matter how long ago you studied the topic in question.
A good solution to this problem is to provide some focus for the interviewee to prepare. If the interviewee knows ahead of time to prepare for statistics questions, then they have a much better shot at the role but more importantly, not to waste the interviewer’s time.
2. Expectation Theory
This leads into the next problem with quant interviews. There is absolutely no expectation of what will happen. You may walk into an almost classroom setting in which you are given a formal exam (e.g. an unnamed bulge-bracket bank’s famous 100 questions in 60 minutes test), a nearly behavioral interview, all the way to one where you must interview along with others (e.g. a certain hedge fund). This lack of expectation and focus leads the candidate to perform poorly and the potential employer to waste their time on an ill-prepared, but most importantly, highly nervous interviewee.
A good solution to this problem would be to provide accurate and fair warning of the types of questions or topic area that will appear. If you want to ask econometrics questions, go ahead, just let the candidate know ahead of time on the topic area and the style of interview (test, panel, group, etc.). For such a wide-ranging field as quantitative finance, this is imperative. It will waste a lot less time for both the interviewer and interviewee.
3. Rapid-firing Leads to Rapid Firing
Speaking of tests, certain companies (mostly prop trading shops and some hedge funds) tend to provide the candidate with an electronic test. There are many variations to this, but it all boils down to performing rapid fire calculations in an absurdly insufficient amount of time. For example, a Chicago prop shop I once interviewed with provided such a test to make rapid fire calculations… 80 of them in under 5 minutes. Here’s the problem with this. It does not test the quality of knowledge (nor the body), only how fast one is performing calculations which can be automated via VBA, C++, or any other given programming language.
One solution, actually provided by a certain other bulge-bracket foreign bank, is to indeed provide a different kind of test. The type of test was actually quite simple in nature. It consisted of precisely four questions, each from a different area of quantitative finance (stochastic calculus, programming, econometrics, undergraduate mathematics). Candidates were given as long as they pleased, and most finished within an hour. The questions were general enough to not be under the brain’s cobwebs, but specific enough to test whether one knew the specific area in question. This was an excellent way of testing the candidates’ body of knowledge than their brains’ computational time. How one applies knowledge is much more useful in business than how fast it is applied (even in high frequency trading).
4. Trust the Candidate Knows What They’re Doing
It is no surprise that to be a good quant, one must be a good programmer. There is no question about that. I’m not quite certain how Silicon Valley tests their candidates, however forcing the candidate to write properly syntaxed code on a piece of paper during an interview setting can be quite daunting. Mistakes are easily made (which would otherwise be caught in a debugger, though unavailable on a piece of paper).
A good solution from the candidate’s point of view is to provide sample code. Verification of whether the candidate indeed wrote the code can be easily achieved by using a code repository such as a GitHub account. A repository with different versions of code can be shared with employers who can see the candidate’s style and skill. The candidate’s coding ability can be verified easily and inexpensively leading to an efficiency gain.
5. Why Not Ask the Why?
One question I’ve never been asked (it is rarely asked in any other field either) is why I want to be a quant in the first place? From my experience, most of the young quants (the ones I encountered in graduate school, both at Rutgers and at other universities) mostly wanted to go into quantitative finance to make a lot of money. The problem with this is the same as doing any other field for the money. Becoming a doctor for the prestige and money is the worst possible reason for becoming a doctor. I also guarantee that if one pursues solely money, they will not only be unhappy, but they will suck at their job. Another problem that arises out of this is that you will not become rich as a quant. There are examples to show the opposite (e.g. Ken Griffin, Cliff Asness, etc), but most quants who go into the field for the money, will not attain that level status.
One solution I suggest, comes from my own pursuit and my own life. I am passionate about trading and math. I actually have fun coming up with new algorithms to trade in new ways. I believe that I am good at it because of my passion which actually drives me to learn more and become even better at it. Without passion and love of the field, whether it is engineering, medicine, law, or quantitative finance, one cannot grow as a person which eventually hurts both the individual and the company.
If interviewers tested for passion, not just pure skill and knowledge, they would find themselves employees who take the company forward in new and innovative ways. After all, anyone can crank out a solution to a partial differential equation, but if they are doing it for their love of math rather than money, they will be much more productive.
If you find my writing helpful and would like to support my endeavors feel free to become a member of mine and get access to ALL of my articles on Medium and other articles for the price of a cup of coffee a month.
Become a member here.