AQR Capital Management Research Software Development Engineer interview questions

If you interview here be prepared to sacrifice a lot of time and effort (and vacation days) for a big unknown. They will not be prompt in communicating with you and will just string you along. If you read through some of the other reviews you will get a sense that I am not the only one with this sentiment. The process started off with a 75 minute hackerrank screening test. They were the type of exercises that weren’t particularly difficult once you actually understood what they were asking. This was challenging to do in the allotted time since they had a lot of extra verbiage. I passed all test cases of the first problem, some of the second and didn’t even get to the third question. I am mentioning this because I received an invitation for technical phone screening even though I did not do that well on the technical pre-screening. The phone screening only lasted about 30-40 minutes. Since my background was in Java, we focused on standard OO questions along with some basic system design. For example, the interviewer wanted me to implement students signing up for courses in a relational database. Although the questions were fairly straightforward, I did not spare any details in answering them. I think the strategy paid off as I was invited to “super-day” almost immediately after the phone screen. Super-day required taking a full day off work which was a little annoying. During the first interview I was asked to briefly describe the system I was currently working on. I got the sense that the interviewer did not want to be there as he was not really paying attention. My guess is that AQR employees get roped into these super-days quite often. The second and third interviews consisted of whiteboard programming questions. The first one was about creating a queuing system to handle multiple users’ tasks and the second one was to implement an algorithm to find the root of a function. I thought I did reasonably well on both. I even think I did better on the second question than the interviewer had expected. All the interviewees were invited by two AQR employees to a local restaurant for lunch which was nice. What wasn’t nice was obnoxiously stressing how working in a bank sucks to a bunch of developers who work in..banks. It definitely crossed the line of a pleasant and appropriate sales pitch for the buy-side. At the time I chalked it up to overeager recruiting. After lunch we again had to implement a task scheduler, this time in the form of a two-hour written test. Mine was not perfect but it got the task done! (sorry) I did not hear for a week after super day but got contacted after following up with HR. They wanted me to either do 2 hour code-pair (again with hackerrank) or come on-site to meet my prospective team. At this point I should have been a little more skeptical but I took their offer in good faith. I came on-site again for a half-day interview figuring they were really serious about me as a candidate. This was a big assumption. The first interviewer started off by saying that I had received “mainly positive” reviews on my previous interview and that his team would continue with “technical grilling”. Everything in quotes were his exact words. I came into it thinking it was going to be an interview but apparently AQR really wanted interrogate me. They made good on their promise and asked me two more whiteboard programming questions, first one on tasks (btw if you haven’t figured it out yet AQR really loves tasks) and the other on maintaining the median of a stream of numbers. I did not do so well on the first question but eventually got through it, albeit in more time than it probably should have taken. Then came the final interviewer. I thought it would be a wind-down session since it was 4:30PM on a Friday but again I was fooled. He started off by asking low-level computer questions which he knew I would be unlikely to answer based on my background. These were definitely not relevant to the job I was applying for. For example, he asked me if the JVM was a register or stack-based VM which no Java/Scala developer will ever have to know. I suspect this was a tactic to establish his intellectual superiority. Finally, he presented to me a coin-flipping thought experiment which was supposed emulate how markets work. I had to show what would happen in the case of market participants (betters/bookies) making bets on the results of an unfair coin flip where nobody actually knows the probability of heads vs tails but the betters think that they do. He was pretty quiet during my explanation so at the end I asked him if he had any other insights about it. He came to a slightly different conclusion because he secretly switched assumptions. Now he was assuming that the bookies knew the probability of the coin flip! I got a little flustered when he tried to forcefully explain this to me (in a pretty disorganized way I might add) and we ended the conversation on kind of a weird note.