XTX Markets Interview ProcessTier 1

ML-Driven Algorithmic Market Maker · London (FX, equities, fixed income, commodities, crypto)

XTX Markets, a London-headquartered electronic market maker, trades almost entirely off statistical and ML models, making it one of the world's largest FX liquidity providers. Its research-heavy interview centers on machine learning, statistics, linear algebra, regression and numerical coding; the online assessment is reportedly one of the industry's hardest single filters. Interviewers probe one or two areas to your limit.

4
hires: Quantitative Researcher, Software Engineer, FPGA / Hardware Engineer, Quant Developer

Online Assessment

XTX Markets is a London-headquartered, machine-learning-driven market maker - reportedly the largest FX spot liquidity provider in the world - and its online assessment is widely reported as one of the hardest single filters in quant hiring, with pass rates often cited around 5-10% (a community estimate, not an official figure). It is the dominant gate in the entire funnel, so it is worth taking seriously: most candidates who get an XTX offer spent the bulk of their preparation here. The format varies by role and by year, and sources genuinely disagree on the exact platform, so the only safe strategy is to prepare for the TOPICS rather than betting on a specific tool. Researchers report a combined machine-learning and probability test, frequently with a small data-analysis component bolted on. A commonly described general or quant version is roughly 20 questions in about 60 minutes spanning probability, linear algebra, calculus, algorithms and brain-teasers, hosted on TestGorilla. Engineering candidates report HackerRank-style timed algorithmic coding under a hard clock. Two details recur across candidate write-ups: the per-question time budget is tight enough that you cannot afford to re-derive the basics under pressure, and you reportedly cannot return to a question once you move past it - so read each one fully, commit, and pace yourself deliberately. Because XTX is a forecasting business operating in extremely low signal-to-noise conditions, the assessment rewards genuine fluency in the statistical and linear-algebra machinery rather than memorized tricks.

Practice this round 15 problems Pro
Quant Researcher
current
QR / ML + probabilityreported TestGorilla / take-home (varies)~60-120 min~20 questions (reported)

For researchers, candidates describe a combined machine-learning, statistics and probability screen, often with a small data-analysis component. Reports consistently mention a tight time limit and an inability to return to a question once you have passed it, so speed on fundamentals matters as much as depth. The representative set below mirrors the topics XTX is known to probe - regularization and the bias-variance tradeoff, PCA and clustering, MLE and estimator derivations, and the cross-entropy / KL machinery behind probabilistic models. Each carries a full worked solution so you can check not just the answer but the reasoning. Drill these until the core derivations (MLE, the OLS normal equations, why we divide by n-1, the first principal component) are automatic, because under a 60-90 second per-question budget there is no time to rebuild them from scratch.

MLE for a Bernoulli ParameterWrite the likelihood, take logs, set the score to zero - the MLE is the sample mean.StatisticsEasySolve →Why We Divide by n-1 in Sample VarianceBessel's correction: the unbiased variance estimator and why the naive divisor underestimates.StatisticsMediumSolve →Cross-Entropy, KL Divergence, and Jensen's InequalityRelate cross-entropy to KL divergence and prove non-negativity via Jensen's inequality.StatisticsMediumSolve →K-Means Clustering From ScratchLloyd's algorithm, the objective it (locally) minimizes, and why initialization matters.Machine LearningMediumSolve →see all 9 problems →
Software & Systems Engineering
current
SWE / engineeringreported HackerRank-style~60-90 min~2-3 problems (reported)

Engineering candidates report HackerRank-style timed algorithmic coding - typically a couple of medium-to-hard problems graded on correctness, edge-case handling, and time and space complexity, with the wall-clock pressure that defines a HackerRank screen. XTX-flavored coding leans toward implementation tasks with a numerical or data-structure twist rather than pure puzzles, and getting the math right (stability, overflow, off-by-one) counts as much as the algorithm. The representative set below spans stable numerical routines, classic dynamic programming, string and big-integer arithmetic, and core data structures - each with a full worked solution. State your complexity up front, confirm the constraints before you type, and prioritize correct, readable code you can fully explain over a terse optimal answer you cannot.

Numerically Stable Log-Sum-Exp and SoftmaxImplement softmax / log-sum-exp without overflow by subtracting the max - a classic XTX-flavored numerical task.CodingMediumSolve →Integer Arithmetic on Decimal StringsAdd / multiply arbitrarily long integers represented as strings, digit by digit with carries.CodingMediumSolve →Edit Distance (Levenshtein Distance)Classic DP over a 2D table; reason about the recurrence and O(mn) time / O(min) space.CodingMediumSolve →Trie Prefix Search ImplementationBuild a trie supporting insert and prefix queries; analyze time and space.CodingMediumSolve →see all 6 problems →
Unlock the full interview funnel

The first stage is on us — you're seeing it in full. Unlock every later round, the must-do problems, and stage-by-stage prep across all 56 firms with QuantVault Pro.

56 firmsEvery stageMust-do problems

Looking for other firms?

Every firm has a curated interview-questions playlist — a ready-to-drill set of problems matched to how that firm interviews, each with a full worked solution.

Practice every firm's OA: timed replicas, free game replicas and guides in one place →

XTX Markets round-by-round guides

XTX Markets interview — FAQ

What is the XTX Markets interview process?

XTX Markets, a London-headquartered electronic market maker, trades almost entirely off statistical and ML models, making it one of the world's largest FX liquidity providers. Its research-heavy interview centers on machine learning, statistics, linear algebra, regression and numerical coding; the online assessment is reportedly one of the industry's hardest single filters. Interviewers probe one or two areas to your limit. The loop runs 6 stages: Online Assessment, Recruiter / First Technical Screen, Take-Home / Technical Data Round (QR track), Onsite — Coding, Onsite — Machine Learning / Research Deep-Dive, Onsite — Probability, Statistics & Linear Algebra.

How many rounds does XTX Markets have?

6 stages in total, starting with the Online Assessment and ending with the Onsite — Probability, Statistics & Linear Algebra.

How do I prepare for the XTX Markets interview?

Work the stage notes above, then drill the XTX Markets interview-questions set and the XTX Markets online-assessment practice — each problem has a full worked solution.