DRW Online Assessment

DRW screens with one of two tracks depending on the role. The Quant Research / Quant Trading track is a timed 45-minute set of about 6 short math and probability problems, all fill-in-the-blank, with the numbers rotated each sitting. The Software Engineer track is a 150-minute, 3-problem coding set (Codility-style, focused on correctness). Either OA feeds the same pipeline: pass the OA, then a 45-minute Zoom round, then an onsite.

QR / QT — 45-min math OA

45-min math OAtimed, 45 min, fill-in~6 of ~15 shown2023 -> 2027

The quant screen is roughly 6 short math / probability problems in 45 minutes, all fill-in (no multiple choice). The recurring families are matrix unknowns, an EV dice game, moments and covariance, a CLT / sampling tail, a Markov-chain stationary distribution, a squircle / Gamma-function area, and (2025+) a matrix-calculus 'invert then differentiate' item. Numbers rotate every sitting, so this timed set drills the methods on the verified problem set. A complete 2027-cycle QR sitting (July 2026, candidate-reported) is replicated question-for-question, including the tournament-bracket graphic.

Software Engineer — SDE 150-min coding OA

SDE 150-min coding OA150 min, 3 problems, no feedback3 problems shown2023 -> 2026

The SDE OA is 150 minutes, 3 problems, with no hidden-test feedback - so test your own edge cases. The recurring problems below are auto-graded.

AI / ML — AI ML coding task

AI ML coding taskcoding, auto-graded1 problem shown

The AI / ML screen is an auto-graded coding task: given an imbalanced binary dataset and a baseline model, write train and test functions that beat the baseline, where only training-side rebalancing reliably moves the score.

What the DRW online assessment screen looks like

On the Quant Research and Quant Trading track the screen is a timed 45-minute form of about 6 fill-in questions with no multiple choice. The numbers rotate every sitting and each question specifies its rounding, so the value you type is graded against the key when you submit. A candidate-reported July 2026 sitting for the 2027 QR intern cycle ran 6 fill-in questions in 45 minutes with an on-screen calculator allowed and webcam proctoring via Honorlock, and one question came with a tournament-bracket graphic. The Software Engineer track is a 150-minute, 3-problem Codility-style coding set with no hidden-test feedback while you work, so you test your own edge cases. The AI / ML task is an auto-graded coding exercise scored against a baseline model.

What the DRW online assessment tests

Across all three tracks the DRW OA rewards one skill: turning a short prompt into an exact number, quickly. Every family reported on the QR / QT math track maps to a topic that also appears among the 87 DRW-tagged problems on QuantVault, so the fastest way to feel the difficulty is to work the closest DRW-reported sibling with a full solution. Each OA family below is paired with a DRW problem in the same topic.

How the OA fits the DRW interview process

The OA is stage one of a six-stage funnel. In the order candidates report it: online assessment; recruiter or phone screen (behavioral plus light technical, sometimes a quick probability or mental-math warm-up); a 45 to 60 minute probability and brain-teaser video round; a trading intuition and market-making round (estimation drills, options basics, game theory); a coding round for engineering roles; and onsite finals. The stage-by-stage map with tips for each round is on the DRW interview process page.

How to prepare for the DRW OA in 7 days

This plan assumes about 90 minutes a day and is written for the QR / QT math track; Day 6 covers the SDE track. Every drill links a QuantVault topic hub with full worked solutions.

DRW OA — FAQ

What is the DRW online assessment?

DRW screens with one of two tracks depending on the role. The Quant Research / Quant Trading track is a timed 45-minute set of about 6 short math and probability problems, all fill-in-the-blank, with the numbers rotated each sitting. The Software Engineer track is a 150-minute, 3-problem coding set (Codility-style, focused on correctness). Either OA feeds the same pipeline: pass the OA, then a 45-minute Zoom round, then an onsite.

What is the format of the DRW OA?

It runs as 3 tracks (45-min math OA, SDE 150-min coding OA, AI ML coding task). QR / QT, Software Engineer, AI / ML, each timed and auto-graded where applicable.

How do I practice for the DRW online assessment?

Work the sample questions above (each with a full worked solution), then take the interactive DRW OA practice on QuantVault. For the full interview, see the DRW interview-questions set.

Is the DRW online assessment automated?

Yes, for every track our candidate reports describe. The QR / QT math OA is a timed fill-in form with the numbers rotated each sitting and no multiple choice, so grading is a numeric check against the key. The SDE coding OA and the AI / ML task are auto-graded, with no hidden-test feedback while the assessment runs. The first human evaluation is the recruiter screen and the 45-minute Zoom round that follow a pass.

How long is the DRW online assessment?

It depends on the track. Quant Research and Quant Trading candidates get 45 minutes for about 6 math and probability problems, roughly 7 minutes each. Software Engineer candidates get 150 minutes for 3 coding problems. Our sources do not report a time limit for the AI / ML coding task; candidates typically describe comparable prop-shop ML screens as a few hours of focused work, but we have not confirmed DRW's limit.

What score do you need to pass the DRW OA?

DRW does not publish a cutoff, and none of our candidate reports include a score. What candidates typically report for comparable prop-shop OAs is that math screens count exactly correct answers with no partial credit, and that the pass line moves with each cycle's applicant pool. With about 6 fill-in questions each miss is a large share of the total, so accuracy on the families you know beats rushing to attempt everything. On the SDE track the auto-grader counts hidden tests passed per problem, so correctness on every edge case beats cleverness.

Does DRW use HackerRank or Codility?

Our candidate reports describe the SDE coding OA as Codility-style: auto-graded, correctness-focused, no hidden-test feedback. We have not verified the vendor from an official DRW source, and DRW lists no public platform for the quant math OA, which candidates describe as a timed fill-in form rather than a coding environment. Prepare for the format (typed exact answers, self-written edge-case tests) rather than for a particular vendor's interface.

What questions are on the DRW online assessment?

On the QR / QT track candidates report about 6 fill-in math and probability problems drawn from recurring families: matrix unknowns, an expected-value dice game, moments and covariance, a CLT or sampling tail, a Markov-chain stationary distribution, a squircle or Gamma-function area, and since 2025 a matrix-calculus invert-then-differentiate item. The SDE track's recurring coding problems (notification cropping, maximum income while already holding one asset, maximum number of equal-sum moves) are listed on this page with auto-graded solutions, and the AI / ML task asks for train and test functions that beat a baseline on an imbalanced binary dataset.