Practice for quant researcher interviews

Researcher interviews split by firm type. Market makers run a fast probability process much like the trader track; systematic funds run a slower one built around a take-home on real data and a defended research discussion. The track below takes probability and statistics first, then research coding and validation, then the quantitative finance the funds ask about.

Sign up freeAlready have an account? Log in

The track

145

items in 9 sections

Starts with
Probability fundamentals
Time to complete
24 h
4 to 6 weeks · 1 h a day

The approach

How to prepare for a quant researcher interview

  1. Step 1

    Find out which process you are in

    Market makers screen researchers much like traders: a fast probability test, then puzzles and games. Systematic funds run a slower process around statistics, Python on real data, a take-home and a research discussion. The firm guide says which, and the two need different preparation.

  2. Step 2

    Statistics is the common ground

    Regression and its assumptions, hypothesis testing, bias and variance and maximum likelihood come up at nearly every firm, so read the statistics lessons on the track and score yourself on their questions. Keep probability sharp beside them: every researcher screen still opens with expected value against a timer, and the researcher bar is higher than the trader one.

  3. Step 3

    Then code the ideas

    The research coding problems ask for pandas and numpy written to pass a code review: summary statistics, labels, autocorrelation, bootstrap intervals. Then the validation course and its problems: leakage, purged cross-validation, walk-forward backtests, multiple testing. The take-home and the project defence are decided on those.

The rules

  • Work from the top

    Read the guide for your firm in the last section, then start at the top. The sections build on each other: probability and mathematics, then statistics, research coding, validation and machine learning.

  • Clear the bar

    A lesson ticks when you complete it, and a coding problem when a submission passes every hidden test. The probability trainer, the course questions and the firm assessment each have a bar, shown as the goal. An article ticks when you open it.

  • Practise from anywhere

    A run, a lesson or a problem counts wherever you start it on Tradermath, not only from this page. Opening an item from the track gives you the session its bar asks for.

  • Free to follow

    Some lessons and coding problems, the probability trainer and every guide are free. The other lessons and problems, and the full firm assessment, need Tradermath Premium. Each item shows its own access.

Sections [9]

  1. 01

    Probability fundamentals

    19 items · 2 h

    Every researcher screen opens with probability, and every later round builds on it. Start from what probability is, then the rules, counting, conditioning and Bayes, and the discrete and continuous distributions up to the central limit theorem.

    Prepares forOnline assessmentTechnical interviews

    1. Fundamentals of Probability and StatisticsCourse · 19 lessons in 6 sections2 h
  2. 02

    Advanced probability

    12 items · 1.5 h

    The tools researcher interviews add on top: expectations of functions and moment-generating functions, Markov chains, martingales and Poisson processes, indicator variables and conditioning tricks, and concentration inequalities. Then drill at speed: the researcher bar is higher than the trader one.

    Prepares forOnline assessmentTechnical interviewsFinal round

    • 11 Lessons
    • 1 Trainer
  3. 03

    Linear algebra and calculus

    7 items · 1.5 h

    The linear algebra and calculus the statistics rests on: portfolios as vectors and covariance as a matrix, least squares as a projection, eigenvalues and principal component analysis, and the calculus of sensitivities, optima and mean reversion. Then a matrix class written from scratch, and the first principal component of a returns panel.

    Prepares forOnline assessmentTechnical interviews

    • 5 Lessons
    • 2 Coding problems
  4. 04

    Statistics and regression

    28 items · 4.5 h

    Regression and its diagnostics, frequentist inference, the multivariate normal, time series, likelihood, simulation and shrinkage. Researcher interviews return to this material twice, in the online test and in the regression round, and each estimator is implemented as soon as it is read.

    Prepares forOnline assessmentTechnical interviewsFinal round

    • 23 Lessons
    • 4 Coding problems
    • 1 Course questions
  5. 05

    Research coding

    19 items · 5 h

    Python on real data, as researcher interviews test it. First pandas written to pass a code review and a solution shaped to be read, then the problems the screens set: summary statistics, labels, point-in-time universes, quote joins and signals.

    Prepares forOnline assessmentTechnical interviewsTake-home

    • 8 Lessons
    • 11 Coding problems
  6. 06

    Validation and backtesting

    28 items · 5.5 h

    The whole validation course, in order, then the same ideas as code. The take-home and the project defence are decided on this material: why research results are false by default, leakage, validation on time-ordered data, multiple testing, reading a backtest, and defending your own research.

    Prepares forTechnical interviewsTake-homeFinal round

    • 21 Lessons
    • 5 Coding problems
    • 1 Course questions
    • 1 Article
  7. 07

    Machine learning

    9 items · 1.5 h

    The machine learning round at the ML-first funds asks about methodology, not libraries: overfitting and the train-test gap, fitting by gradient descent, choosing a model family on noisy data, trees, bagging and boosting, building features without leaking, and neural networks in brief.

    Prepares forTechnical interviews

    • 8 Lessons
    • 1 Coding problem
  8. 08

    Quantitative finance

    16 items · 2 h

    The finance the funds ask about: returns, the Sharpe ratio, mean-variance and diversification, the efficient frontier, risk parity and stress testing, factor models, the cost of carry and tail risk, then where a statistical edge comes from and how the market it trades in is structured.

    Prepares forTechnical interviews

    • 15 Lessons
    • 1 Course questions
  9. 09

    Firm guides and assessment

    7 items · 48 min

    The firm assessment nearest to a researcher screen, at its real length, and the process guides for the funds with a research discussion round. A full run needs Tradermath Premium; a free account can sit a shorter example.

    Prepares forOnline assessmentTechnical interviewsFinal round

    • 6 Articles
    • 1 Assessment

Interview ready

How quant researcher interviews run

The stages in the order firms run them. Select a stage for its rounds and the track sections that prepare them.

Stage 02 of 05

Timed online assessment

Probability and statistics, often with a coding or data component. At the speed-driven firms this is where reported rejections concentrate, and speed matters more than knowledge.

The roundsPrepared by

Full process The Quant Researcher Interview Process

Then zoom in on a firm

The track gets you to level in general. Each firm runs its own version of the process, and its page has the assessments, questions and guide for it.

Quant researcher interview preparation: questions

How do quant researcher interviews differ from trader interviews?
Researchers get fewer mental-math sprints and market-making games, and more statistical inference, regression, machine learning methodology and a defended research discussion. Jane Street and DRW blur the line with hybrid seats; systematic funds such as D.E. Shaw and Two Sigma keep the two tracks clearly separate.
What do quant researcher interviews ask?
Probability and expected value at every firm and every stage, then statistics: OLS and its assumptions, hypothesis testing, bias and variance, maximum likelihood. Systematic funds add Python on real data, a take-home data challenge and a research discussion where your own project is taken apart, assumption by assumption. ML-first firms add a round on regularisation, ensembles and generalisation. Market makers stay closer to the trader process: puzzles and betting games.
How much coding is in a quant researcher interview?
Less than a developer interview, more than a trader one. The online test often has a short coding or data section, the technical rounds ask pandas and numpy on real data at systematic funds (one large fund asks for an efficient linear regression in numpy), and the take-home is a piece of research code end to end. The track's research coding and validation problems run in the workspace against hidden tests, in Python.
Do I need a PhD?
Not everywhere, but the research discussion round is weighted far more heavily for PhD candidates, and systematic funds hire mostly at that level. Masters and undergraduate candidates see more of the probability, statistics and coding rounds and less project defence.
How long does the Quantitative Researcher Role Track take?
Four to six weeks at an hour a day. The track counts about 24 hours of sit-down time, and that figure is a floor: a coding problem usually takes more than one attempt, and clearing a bar takes a few tries. The validation course is the core of the track and is worth reading in order; the coding problems beside it are where the same ideas get implemented.
Is there a machine learning round?
At some firms. XTX Markets, G-Research, Two Sigma, Jump Trading and Point72 Cubist run a distinct round on regularisation, ensembles, neural networks and generalisation in practice; market makers mostly do not. The track has a machine learning section on exactly those topics, with the mathematics and statistics it rests on earlier in the track, and the firm guides describe what each of those rounds asks.
How is progress tracked?
From what you have already done: a lesson counts when you complete it, a coding problem when a submission passes every hidden test or you mark it complete from the problem page, and a trainer or firm assessment once your best run clears its bar. A full firm assessment needs Tradermath Premium. An article or firm guide counts once you have opened it while signed in.