Quantitative Researcher

If I had gone the distance, the full interview process would involve 3 phone interviews and then an on-site interview in New York. I'd love to tell you about the on-site interview but I didn't make it that far. Practice your probability brainteasers! Statistics buffs with some coding experience are probably ideal for this role's interview process.

The first phone interview had a pretty straightforward algebra brainteaser, a test to see if you can take given information and formulate a simple series of equations and solve them. Nothing too elaborate, but since you're on the phone it can be a bit stressful. Then there was a game theory question, you and our opponent have allowed moves and you have to come up with both players' strategies to maximize their expected returns. I couldn't get a proper expression, but came up with the right qualitative reasoning so I guess I got a pass.

Second phone interview was a coding test (Python) through coderpad. Link was provided at the start of the interview so there wasn't time to get experience with the interface. Fortunately I had the chance to practice because of an earlier interview with Citadel, who gave out the link several days in advance. The interviewer gives you the problem statement and any hints you might need, and you go ahead and write out functioning code. The coderpad website has a Python interpreter (among other languages) so you and your interviewer can run code and test it against some edge cases. Technically coderpad has code autocompletion but I found it somewhat buggy; I strongly recommend you practice some simple problems (e.g. Project Euler or Hackerrank) without relying on autocompletion. I didn't quite finish the code before time ran out but apparently that's not necessary to advance to the third round phone interview.

The third round interview was probability brainteasers, a reasonably generic one involving coin tosses (or darts thrown at a board, or sandbags thrown in a pit, etc). I was asked not to share problems, but Glassdoor and some other websites have very similar problems. Unfortunately, when asked about my intuition about various outcomes, I guessed wrong and took a lot of hints to get to the right answer. Then I was asked to qualitatively explain the correct answer, which I did poorly because I was still wrapping my head around it.

Overall I suspect that my questionable performance on the game theory and probability brainteasers from the first and third interviews ended my candidacy. Jane Street people are famously (notoriously?) statistics, betting, and market-making oriented so these kinds of problems are very important to them. For example, I've heard that some entry-level JS employees (especially Traders?) are required to play poker, then scored on their analysis of different hands and their P&L. My interview advice to any company is 1. trawl the internet for their typical kinds of interview questions and prepare accordingly, and 2. try to get the name of your interviewer and stalk their LinkedIn profile. Software engineers give coding and algorithms questions, otherwise you're getting probability brainteasers and data analysis questions if you put "machine learning" anywhere on your resume.

Notably absent from the interview process (at least on the phone) were any serious mental arithmetic questions or some of the more bizarre "make a bet on the correctness of the answer you just gave" things I've heard rumors about. That may have a lot to do with the position, Quantitative Researchers there are coding and data analysis focused and therefore probably tamer, versus Traders who seem to live and breathe bets and rapid-fire arithmetic.