Citadel interview notes, page 9 of 14

Large Enterprises

Most of these Citadel write-ups come from Software Engineers (106), then Data Scientists (33) and Machine Learning Engineers (1). The most reported rounds are the technical phone screen (111), the system design round (16), the onsite coding round (4) and the behavioral / leadership round (3).

140 interview notes · updated Oct 2026

Citadel·Software Engineer·Onsite - System Design / Architecture

Jun 2026

Citadel system design question focused on shared state across processes on a single machine. Short but surprisingly tricky if you haven't thought about IPC much.

You have a counter running on a single host, and multiple separate processes need to read and update it. How do you design this so it works correctly across process boundaries?

“My first instinct was shared memory with a mutex and I said it pretty quickly, but then they pushed on what happens when a process crashes mid-write.”

View Post

Citadel·Data Scientist·Technical Phone Screen

Jun 2026

Citadel data scientist interview that leaned heavily into PhD background, which I wasn't fully ready for. The questions were more reflective and self-critical than I expected from a quant shop. Less coding, more 'tell me how you think about research.'

Which two courses from your PhD shaped how you approach empirical modeling, and why those specifically?
Walk me through a research project where your first approach didn't work. What changed after getting feedback, and how did you measure whether the new direction actually improved things?
If your advisor and a collaborator were each asked to name one thing you should improve, what would they say, and what have you actually done about it?

“I named two courses but fumbled explaining the 'why' in any concrete way.”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Citadel software engineer interview with a string manipulation problem that's trickier than it looks. The palindrome constraint combined with the periodicity constraint means you have to think carefully about which character positions are actually linked together.

Given a string of length N and an integer k where k divides N evenly, find the minimum number of character substitutions needed so that the resulting string is both a palindrome and k-periodic (meaning every character equals the one k positions ahead of it).

“The two constraints interact in a non-obvious way.”

View Post

Citadel·Data Scientist·Technical Phone Screen

Jun 2026

Citadel data scientist interview with a deep dive into OLS diagnostics. One question but it had a lot of moving parts and I definitely didn't nail all of them.

In OLS regression, how do you define outliers, high-leverage points, and influential observations? Walk through how Cook's distance is derived, how it connects to leverage and studentized residuals, how you'd build a diagnostic workflow to detect each type, what happens to your conclusions if you drop the most influential point, what robust alternatives exist, and how you'd justify any decisions about these points in a model review.

“This question is basically three questions stapled together and I treated it that way, which was probably a mistake.”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Citadel software engineer interview that centered on array merging, starting from the classic two-array problem and then pushing into a three-array variant with deduplication. The algorithmic depth ramps up faster than you'd expect from what looks like a warmup question.

Given two sorted arrays where the first has extra capacity at the end to hold both arrays' elements, merge the second into the first in-place and return it sorted. Solve it in O(m+n) time.
Extend the previous problem to merge three sorted arrays and remove any duplicates so the final result is strictly increasing.
Compare using a min-heap for k-way merging versus sequentially merging two arrays at a time. How does deduplication interact with each strategy?

“The two-pointer-from-the-end trick is the whole point here.”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Citadel SWE interview with three medium-level coding problems. The twist on the first one tripped me up a bit since it looked familiar but had a uniqueness constraint I almost missed.

Given a string or collection, find all substrings matching some condition, but the final result must contain only unique substrings (no duplicates).

“Looked like a problem I'd seen before and almost just coded the standard approach without thinking about uniqueness.”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Citadel quant engineer interview that went deep into regression theory. The core question was about deriving OLS slopes and showing their product equals r squared. More math-heavy than I expected for an engineering role.

Derive the least squares slopes for regressing y on x and for regressing x on y. Then show that multiplying the two slopes together gives you the squared Pearson correlation coefficient.
Why are the two regression lines not inverses of each other? What does that tell you geometrically about what each regression is actually minimizing?

“I got the individual slope formulas fine, Cov(x,y)/Var(x) and Cov(x,y)/Var(y).”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Citadel coding round, one algorithmic problem on tree diameter. Pretty focused session, no fluff.

Given an undirected tree, find the longest path (diameter) and return the actual list of nodes along that path.

“The two-BFS trick is something I'd seen before but blanked on the second pass.”

View Post

Citadel·Software Engineer·Technical Phone Screen

Jun 2026

Quant-style probability question from Citadel, probably a phone screen or early technical round. Clean problem, straightforward if you know linearity of expectation, but easy to fumble the combinatorics under pressure.

You have 6 red cards and 4 black cards shuffled randomly into a line of 10. What is the expected number of adjacent pairs that share the same color?

“The key move is recognizing you don't need to think about the full distribution, just treat each of the 9 adjacent slots independently.”

View Post

Citadel·Data Scientist·Technical Phone Screen

Jun 2026

Citadel data scientist interview, technical screen with a stats/optimization question. Pretty math-heavy and they wanted a full derivation, not just the answer.

You have a set of observed values and you want to predict a single constant for all of them. If you minimize the sum of absolute errors, what constant should you pick? Derive it and discuss whether the answer is always unique.

“I knew the answer was the median but fumbled the actual derivation under pressure.”

View Post