Citadel Super Day ML / Statistics Interview

The Citadel Super Day ML / Statistics interview is stage 7 of 8 in the Citadel interview process, after the Super Day Coding. Format: 60 min · whiteboard. A Super Day round for QR/ML tracks: regression, model diagnostics, regularization.

What the Super Day ML / Statistics tests

They probe assumptions hard - be ready to justify 'why does this hold.'

QuantVault maps 22 practice problems to this round. By topic, the biggest groups are regression (10), time series (5), statistics (4), linear algebra (2); by difficulty, 0 easy, 11 medium and 11 hard. 22 of them were reported by candidates who interviewed at Citadel, most recently in Jun 2026.

• ML and statistics are part of the three to five hour-long quant research onsite interviews.

Who takes this round: Quant Researcher candidates. See the full Citadel process and who takes each round.

How to prepare for the Super Day ML / Statistics

• Know OLS cold, then its assumptions. Be ready to state Gauss-Markov, derive β̂ = (XᵀX)⁻¹Xᵀy, and explain what breaks under heteroskedasticity / collinearity / endogeneity.
• Regularization with a reason. Ridge vs. lasso, the bias-variance trade-off, and how you'd pick λ (CV vs. MSE minimization) — they probe 'why does this hold' hard.

Further reading: How quant superdays work · Behavioral questions at trading firms · Market-making game interview guide.

Question types in this round

The 5 largest question types in the Citadel Super Day ML / Statistics practice set; across them the most common sub-topics are matrix operations, variance, regularization, convex optimization, eigenvalues, hypothesis testing. The questions themselves, with worked solutions, are in the round practice.

Common mistakes

• Diagnose before you model. Residual plots, leverage, multiple-testing / overfitting traps. Show you'd validate out-of-sample, not just fit.

Prepare with QuantVault

The same round at similar firms

Compare with similar firms