Last year I interviewed at CitSec and some other firms for a role as a Quantitative Developer.

The process was difficult, but I wrote down everything that I was asked so that I can better hone my skills and assess where I need to improve.

Interview Prep Notes – Compiled by CJ

Algo · Lower Level (C++/OS/HW) · Quant Dev

About these notes

Answers, where they exist, are abridged and may be inaccurate. Each section contains:

• Conceptual questions – questions I was explicitly asked.

• Coding questions – questions I was explicitly asked.

• Extras – additional questions and coverage checks used during prep.

Questions may not be representative of generic interviews. They are likely tailored to my background.

Quant

Coding Questions

Case-study style questions

Obscured algo questions, with lots of prompting for requirements and context before implementation.

• Many places: Implement a simplified matching engine. Similar to LeetCode 1801 – Number of Orders in the Backlog, often with twists.

  • IMC has a fairly interesting twist, asked in its final round. See this problem.

• k-sorted array.

• Optimal allocation of batched fills. Roughly, you have internal buy and sell client orders that are executed externally every N seconds. On each fill, you want to distribute the filled quantities to the internal clients so that it is “fair” and their orders are maximally filled. This was an open-ended, case-study-style question, with lots of discussion about what constitutes “fairness,” constraints, and so on. The best place to start is by implementing an orderbook, after which you can take a deeper dive into orders by testing your ability to identify iceberg orders.

• Implement a very simplified backtest system. You get a stream of TOB data with capture and transact time. The strategy only has one outstanding order at a time. I cannot find the exact question, but doing this one will give you the background you need to understand how to handle incremental and snapshot updates.

• Auction matching system. Given a list of bids and asks at different prices, choose a price that maximizes execution volume. I could not find this problem anywhere, so I created it myself here.

• Travelling salesman problem. A common problem in computer science. It was this one exactly.

Pure algo style questions

Miscellaneous places, 2024

• Maximum Number of Visible Points – a simplified version upon prompting where we did not need to handle polar-coordinate coordination, and we could assume our position was centered. You can also do this one, which I was asked about at the CME, to solidify your knowledge.

• Sliding Window Maximum.

• Extended or modified versions of:

  • Design a Food Rating System.

  • Stock Price Fluctuation.

Miscellaneo