Preparation library

Quantitative researcher interview preparation

Practice inference, model validation, algorithms and research communication.

Reviewed 8 September 2026 · Original QuantPrep guide

Make the experiment defensible

A research answer connects an estimand, a dataset, a validation design and a decision. Calculating a coefficient matters; explaining why it is identifiable and honestly evaluated matters more than quoting a formula.

Two Sigma’s guidance includes open-ended analysis, coding, algorithms and statistics or research-domain discussion. Citadel’s campus guidance also emphasizes programming and explaining trade-offs.

Worked example: effective information

For 240 observations with AR(1) correlation 0.5, the large-sample approximation n_eff = n(1−ρ)/(1+ρ) gives 80. This is an approximation under a specified dependence structure, not permission to substitute 80 in every statistical test. Overlapping labels or regime changes may invalidate the model.

Preparation order

  1. Probability, expectation, variance and conditioning.
  2. Inference, regression, matrix operations and estimator uncertainty.
  3. Leakage, multiple testing, time-aware splits and costs.
  4. Algorithms, numerical stability and clear research exposition.

Bring one real project

Explain its baseline, failed experiments, validation choices, limitations and next experiment. Implement analysis outside QuantPrep too: numeric drills and self-evaluated prompts do not provide a dataset workspace or executable coding sandbox.

Try a fresh question

This optional exercise uses the actual parameterized bank.

Keep the practice connected

Train Data & Modeling · Follow the Quant Path · Evidence standard