Marshall Wace Internship & Interview Questions Guide

The Marshall Wace interview process for quantitative roles runs across four to five stages, according to Marshall Wace's Quant Internship Application Guide and Quantitative Research Internship Process page. You start with initial technical assessments (either an in-person test at a local center or three sequential online tests with a 168-hour limit for each), followed by a 30-minute Human Capital Zoom screen and a 60-minute technical project presentation. Finalists attend a London onsite day featuring a 50-minute brainteaser interview, a programming test, and manager meetings, while Fundamental Analyst (MW Elevate) candidates are sourced exclusively through dedicated headhunting agency partners. Soreno, an AI voice interview practice app for finance and banking candidates, includes free paper LBO and valuation drills with no sign-up, and voice mock interviews that test your accounting, valuation, and fit answers out loud.

Updated

  • Stage 1 (Online Route). Three sequential tests. First is a 60-minute Codility programming test. Passing that unlocks a 60-minute Technical exam on Correlation-One. Passing that unlocks a 45-minute Data exam on Correlation-One. Each link expires automatically 168 hours after it is sent.
  • Stage 1 (In-Person Route). A proctored in-person assessment at a test center local to you, which cannot be administered virtually.
  • Stage 2 (Human Capital Interview). A 30-minute Zoom call with a member of the Human Capital (HC) team covering why you want the internship, how it fits your career goals, and competency examples that reflect Marshall Wace values.
  • Stage 3 (Project Presentation Interview). A 60-minute Zoom interview. You present a technical project you previously worked on (from a university assignment, PhD thesis, or personal project) for around 15 to 20 minutes, leaving the rest of the time for technical questions.
  • Stage 4 (Onsite Interviews in London). Includes a 50-minute Brainteaser interview where you solve two mathematical brainteasers (plus a Programming test on the in-person process track), informal Management interviews where you meet two or three managers at a time, and a shadowing session with a member of the Quant teams.

Key facts, benchmarks, and worked examples

Below is a breakdown of marshall wace interview, sourced from official firm, school, and regulatory documentation.

How the Marshall Wace interview process works4 rows

Marshall Wace manages quantitative, systematic, and fundamental strategies, specializing predominantly in long/short equity. In 2002, the firm launched MW TOPS, the world's first Alpha Capture application. Because the firm splits its work across systematic research, high-performance engineering, and fundamental stock analysis, the interview stages depend on the track you choose.

Marshall Wace programs, locations, and 2027 dates (Marshall Wace careers pages)

Marshall Wace programs, locations, and 2027 dates (Marshall Wace careers pages)
ProgramLocationsKey Dates / DurationEligibility & Application Route
Quantitative Research InternshipLondonJune 28 to September 3, 2027 (fixed). Applications close January 12 at 12pm GMT.Penultimate-year STEM Master's or PhD (2028 graduation). Direct application.
Technology InternshipLondon, New York, Singapore, Hong KongLDN: June 28 to Sept 3, 2027. NYC: June 14 to Aug 22, 2027. SG: May 17 to July 25, 2027. HK: May to Aug 2027.All students regardless of level or area of study. Direct application to one office only.
Quantitative Associate ProgramLondonFull-time permanent cohortMinimum Master's or PhD in STEM, or 2 to 3 years of relevant experience. Direct application.
MW Elevate (Fundamental Analyst)London, New York, Hong Kong, Singapore, Shanghai2-year full-time program2 to 4 years total experience (minimum 1 year analyzing companies). Agency headhunters only.
  1. Part 1

    Before you apply, know the (1)

    Before you apply, know the single-application rule. The firm can only process one application per candidate each season, and applying to multiple roles or international offices slows down the review of your submission. A cover letter is not necessary, and every resume is reviewed by a human. Under its Applicant Privacy Notice, Marshall Wace also uses artificial intelligence to process information from CVs and job application forms and to transcribe or summarize audio from interview calls, though hiring decisions are not based solely on automated decision-making.

  2. Part 2

    If you want to warm (2)

    If you want to warm up on valuation, mental math, or behavioral stories before your first screen, you can practice a case free, no sign-up.

Coding challenges and technical exams: Codility and Correlation-One2 points

If you are invited to the three-part online testing suite, you must complete every test independently without search engines or large language model chatbots. If you require extra time for a condition such as dyslexia, you must declare it on your application form or notify the recruitment team before opening the test links so they can issue extra-time versions.

  1. Part 1

    The first assessment is the (1)

    The first assessment is the 60-minute Codility programming test. It checks basic coding techniques and your ability to write and use simple functions rather than advanced language-specific features. Codility typically supports C, C++, C#, Go, Java, JavaScript, Python, Ruby, and Scala. Marshall Wace permits you to write your code in your own local IDE and paste your finished answers into Codility, provided you do not use LLMs. To prepare for Python or C++ coding challenges, disable any AI autocomplete plugins in your local IDE beforehand and practice writing clean, modular functions from scratch that handle edge cases, array sorting, and basic data structures without external libraries.

  2. Part 2

    The second assessment is the (2)

    The second assessment is the 60-minute Technical Exam on Correlation-One. It tests probability theory, statistics, regression analysis, algebra, matrices, optimization, and algorithm design. The third assessment is the 45-minute Data Exam on Correlation-One, which evaluates light quant research. You download several CSV files (requiring about 10MB of local storage space) and answer analytical questions using your own local data manipulation environment, such as Python, MATLAB, or R. Because you only have 45 minutes, have a local Python script or Jupyter notebook open before starting the timer so you can immediately load the CSVs into pandas, inspect missing values or outliers, merge tables, and compute summary statistics and regressions.

Official Marshall Wace sample questions and our worked solutions10 points

In the Appendix of its application guide, Marshall Wace publishes five retired questions from previous iterations of the Quant Associate Program (four from the Stage 1 Technical Test and one from the Stage 4 Onsite Brainteaser interview). The official PDF gives the questions and point values without answers. Below are the exact official prompts followed by step-by-step solutions worked out by our team.

  1. Part 1

    Technical Test Question 1 [3 (1)

    Technical Test Question 1 [3 marks]: A researcher takes a random sample of 25 fourteen-year-old students from a large population and gives them an IQ test. The population mean IQ is known to be 100. The first student is found to have an IQ of 150. What is your expectation for the average IQ for the sample of 25 students?

  2. Part 2

    Worked solution by our team: (2)

    Worked solution by our team: The sample consists of 25 independent draws from a large population with mean 100. The first student's score is fixed at 150. Each of the remaining 24 students is drawn randomly from the population, so the expected IQ for each of the remaining 24 students is 100. By linearity of expectation, the expected total sum of the 25 scores is 150 + (24 x 100) = 2,550. Divide that expected total by 25 students: 2,550 / 25 = 102.

  3. Part 3

    Technical Test Question 2 [3 (3)

    Technical Test Question 2 [3 marks]: You have a table of items containing three columns of length n. Columns A and B contain only integers, while Column C contains only strings, all of which are 6 characters long. You use quicksort to sort column A. What is the average time complexity when sorting n elements?

  4. Part 4

    Worked solution by our team: (4)

    Worked solution by our team: Sorting n integer elements using quicksort requires O(n log n) comparisons on average. Columns B and C do not increase the comparison complexity because comparing two integers takes O(1) time and swapping rows or pointers with fixed 6-character strings also takes O(1) time per operation. The average time complexity is O(n log n).

  5. Part 5

    Technical Test Question 3 [3 (5)

    Technical Test Question 3 [3 marks]: Find all real solutions to (x^2 + 7x + 11)^(x^2 + x − 30) = 1.

  6. Part 6

    Worked solution by our team: (6)

    Worked solution by our team: A real expression of the form A^B equals 1 in three cases. Case 1: The base A = 1. Set x^2 + 7x + 11 = 1, or (x + 2)(x + 5) = 0, giving x = -2 and x = -5. Case 2: The exponent B = 0 and the base A is not 0. Set x^2 + x − 30 = 0, or (x + 6)(x − 5) = 0, giving x = -6 (where A = 5) and x = 5 (where A = 71). Both are valid. Case 3: The base A = -1 and the exponent B is an even integer. Set x^2 + 7x + 11 = -1, or (x + 3)(x + 4) = 0, giving x = -3 (where B = -24, even) and x = -4 (where B = -18, even). Collecting all three cases yields six solutions: x = -6, -5, -4, -3, -2, 5.

  7. Part 7

    Technical Test Question 4 [5 (7)

    Technical Test Question 4 [5 marks]: A, B, C, D, E, F, G, H are distinct digits within the range 0-9 (inclusive). Work out their values such that the following two arithmetic equations hold: AA x BB = CDE and BB / AA = F / GH.

  8. Part 8

    Worked solution by our team: (8)

    Worked solution by our team: Since AA = 11 x A and BB = 11 x B, the product is AA x BB = 121 x A x B = CDE (< 1,000), so A x B <= 8. Also, BB / AA = B / A = F / GH, where F is a single digit and GH is a two-digit number (>= 10), which forces B < A. Testing distinct non-zero digit pairs with A x B <= 8 where 121 x A x B produces three distinct new digits: A = 4 and B = 2 gives A x B = 8 and CDE = 968. Then B / A = 2 / 4 = 1 / 2 = F / GH. Using the remaining unused digits {0, 1, 3, 5, 7}, set F = 5 and GH = 10. All eight digits are distinct: A = 4, B = 2, C = 9, D = 6, E = 8, F = 5, G = 1, H = 0 (44 x 22 = 968 and 22 / 44 = 5 / 10).

  9. Part 9

    Stage 4 Onsite Brainteaser Question: (9)

    Stage 4 Onsite Brainteaser Question: We each take turns to roll a dice. Each time a number comes up which had not previously come up, we cross it out from the list of numbers 1, 2, 3, 4, 5, 6. The winner is the player to cross out the last number. Would you prefer to play first or second?

  10. Part 10

    Worked solution by our team: (10)

    Worked solution by our team: Let P(k) be the probability that the player whose turn it is to roll wins when k uncrossed numbers remain. On that turn, the player rolls a new number with probability p_k = k/6 (moving the game to state k − 1 for the opponent, where the opponent wins with probability P(k − 1), and where P(0) = 0 since hitting the last number at k = 1 wins immediately) or rolls an already-crossed number with probability 1 − p_k (handing state k to the opponent). Thus P(k) = p_k x (1 − P(k − 1)) + (1 − p_k) x (1 − P(k)), which simplifies to P(k) = (6 − k x P(k − 1)) / (12 − k). Computing from k = 1 to 5: P(1) = 6/11, P(2) = 27/55, P(3) = 83/165, P(4) = 329/660, and P(5) = 463/924. At the start (k = 6), the first roll always crosses out a number (p_6 = 1), handing state k = 5 to the second player. The first player wins with probability P(6) = 1 − P(5) = 461/924 (about 49.89%) and the second player wins with probability 463/924 (about 50.11%). Choose to play second.

Alpha Capture (MW TOPS), Quant Tracks, and MW Elevate2 points

When you interview at Marshall Wace, interviewers evaluate where you fit across its systematic, engineering, or fundamental teams. In 2002, Marshall Wace launched MW TOPS, the world's first Alpha Capture application. In an alpha capture system, fundamental long/short equity trade ideas and return forecasts are collected and systematically evaluated alongside quantitative signals. Quantitative researchers and developers then build models and portfolio construction algorithms that weight those ideas, forecast asset returns across horizons from hours to years, and optimize live execution to reduce risk and trading costs while maximizing returns. Within the Quantitative Associate Program in London, candidates apply to either Quant Research or Quant Development (part of the Quantitative Implementation team). In Technology Internships, interns work in pairs reporting to a senior engineer on projects such as a backtesting framework for the quantitative trading system, horizontally distributed autonomous coding agents, or an AI-driven user behavior analytics platform. We also note that Marshall Wace offers 4- and 6-month UK placements for MEng Computing and Maths & Computing students at Imperial College London, advertised exclusively through the university department. For candidates targeting a fundamental long/short equity role through MW Elevate, the two-year global Fundamental Analyst Program across London, New York, Hong Kong, Singapore, and Shanghai, Marshall Wace does not accept direct applications on its website. Instead, dedicated recruitment agency partners proactively headhunt and screen candidates with 2 to 4 years of work experience (including at least 1 year analyzing companies). Analysts join a specific sector or regional Fundamental Investment team with training in Risk, Trading, Technology, and Quantamental Research (and case-by-case flexibility to rotate into a second team). Under New York City's Pay Transparency Law, Marshall Wace lists the annual base salary for US-based MW Elevate Fundamental Analysts at $150,000, plus eligibility for a year-end discretionary bonus. Because MW Elevate requires at least one year of experience analyzing companies and looks for candidates who think commercially and challenge consensus, you should prepare both a long and a short equity valuation pitch before speaking with headhunters or portfolio managers. A clear fundamental long/short pitch covers four parts: a differentiated thesis showing where your revenue or margin forecast diverges from consensus, a valuation bridge (such as a discounted cash flow or trading multiples model) quantifying the upside or downside, the specific catalyst that will cause the market to re-rate the stock, and the primary sector, liquidity, or execution risks that could invalidate your view.

  1. 1. Quant Research Track

    Quant Research Track

    Researchers conceptualize, backtest, and analyze models forecasting asset returns on horizons of hours to years, using Python, R, or MATLAB to reduce risk and cost while maximizing returns.

  2. 2. Quant Development (Quantitative

    Quant Development (Quantitative Implementation) Track

    Requires a minimum Master's degree in STEM (with up to 3 years of experience welcomed) and expert-level programming in Python, MATLAB, C++, or Java alongside probability, statistics, linear algebra, and numerical methods. Developers build production alpha algorithms, portfolio construction processes, low-latency Linux pipelines, and CI/CD monitoring.

From , , , , , , and .

Quant Research and Quant Associate stages, round by round

Marshall Wace publishes two official descriptions of the Quantitative Research Internship assessment pipeline: the downloadable application guide PDF and the Quantitative Research Internship Process webpage. Both lead to the same middle and final interview stages, and differ only in how Stage 1 testing is delivered. On the Quantitative Research Internship Process webpage, Stage 1 is listed as an in-person test administered at a testing center local to you (beginning from September 21, after applications open on August 17, 2026, with positions expected to be filled by December), with an onsite Programming test included on the final interview day. In the downloadable application guide PDF (where applications close October 22 at 17:00 BST, Human Capital calls run September 22 to November 7, video interviews run October 6 to November 14, and onsite interviews run October 20 to November 21), Stage 1 is a sequence of three online tests (Codility, Technical, and Data), where each test is triggered only after you pass the previous one. Check your invitation email to confirm whether your cohort and location are assigned the local testing center or the three-part online sequence.

  • Stage 1 (Online Route)

    Three sequential tests. First is a 60-minute Codility programming test. Passing that unlocks a 60-minute Technical exam on Correlation-One. Passing that unlocks a 45-minute Data exam on Correlation-One. Each link expires automatically 168 hours after it is sent.

  • Stage 1 (In-Person Route)

    A proctored in-person assessment at a test center local to you, which cannot be administered virtually.

  • Stage 2 (Human Capital Interview)

    A 30-minute Zoom call with a member of the Human Capital (HC) team covering why you want the internship, how it fits your career goals, and competency examples that reflect Marshall Wace values.

  • Stage 3 (Project Presentation Interview)

    A 60-minute Zoom interview. You present a technical project you previously worked on (from a university assignment, PhD thesis, or personal project) for around 15 to 20 minutes, leaving the rest of the time for technical questions.

  • Stage 4 (Onsite Interviews in London)

    Includes a 50-minute Brainteaser interview where you solve two mathematical brainteasers (plus a Programming test on the in-person process track), informal Management interviews where you meet two or three managers at a time, and a shadowing session with a member of the Quant teams.

  • Offers and Feedback

    Marshall Wace aims to share stage feedback within 10 working days (up to 2 weeks) and delivers final offers via a Zoom call with the Human Capital team within two weeks of the management interviews, including your specific team allocation.

Official sources: , , , , and .

1 practice question and sample answer

The practice prompts and sample answers below are Soreno's own, written for invented candidates.

Practice question and sample answer1 question

Follows the exact three-part prompt from Marshall Wace's Stage 3 guide (where the idea came from, what you did, and how it worked), uses 'I' instead of 'we' to establish personal code ownership, and proactively names a concrete modeling flaw before the interviewer asks.

  1. Sample Stage 3 project presentation answer: Walk us through a technical project you worked on, where the idea came from, what you personally built, and one limitation you had to solve.

    Show the sample answer

    Sample answer. Example answer written by our team. The applicant is invented. For my Master's thesis, I investigated whether order-book imbalance at 5-minute intervals could improve short-horizon return forecasts on European liquid equities once transaction costs were included. The idea came from noticing that standard linear volume-imbalance regressions looked predictive in-sample but decayed sharply during high-volatility open and close auctions. I built the ingestion and backtesting pipeline in Python from raw tick data, aligning trades and quotes and stripping out auction prints. Initially, my ordinary least squares regression overfit to extreme volume spikes, which inflated turnover. I switched to ridge regression with an L2 penalty tuned via walk-forward cross-validation and added a turnover penalty directly into the position-sizing rule. Across our out-of-sample test set, net risk-adjusted returns improved while daily turnover dropped by a third. The biggest limitation remaining in my model is that it assumes constant mid-to-bid-ask spread costs rather than modeling nonlinear market impact for larger order sizes, which is the first part I would rebuild next.

    Follows the exact three-part prompt from Marshall Wace's Stage 3 guide (where the idea came from, what you did, and how it worked), uses 'I' instead of 'we' to establish personal code ownership, and proactively names a concrete modeling flaw before the interviewer asks.

How to prepare for the Stage 3 Project Presentation and Stage 2 HC screen

How to prepare for the Stage 3 Project Presentation and Stage 2 HC screen

Marshall Wace states that you do not need to spend massive amounts of time cramming new topics for the technical interviews, because the questions are built to probe your existing knowledge rather than what you revised the night before. Focus your preparation on structuring your 60-minute Stage 3 Project Presentation and sharpening your Stage 2 stories with behavioral interview practice. Use this checklist written by our team to prepare your Stage 3 presentation and local test environment before your links arrive: When you are ready to rehearse your technical walkthroughs, valuation pitches, and behavioral stories out loud with an AI interviewer that scores your answers, start practicing with Soreno today.

  1. Point 1

    Set up your local data (1)

    Set up your local data and coding environment before opening Stage 1 links: Verify that your preferred IDE and Python, R, or MATLAB setup are ready with at least 10MB of free disk space for the Correlation-One Data Exam CSV files.

  2. 2. Pick one technical

    Pick one technical project you own end to end for Stage 3

    Choose a PhD chapter, university assignment, or personal project where you wrote the code and chose the mathematical methods yourself.

  3. 3. Time your slide

    Time your slide deck to around 15 to 20 minutes

    Cover where the idea came from, what you personally did, and how the model or system worked, leaving the rest of the 60-minute interview open for interviewer questions.

  4. 4. Audit every assumption

    Audit every assumption and design trade-off in your project

    Be ready to explain why you chose a specific regression, optimization constraint, or data structure over the alternatives.

  5. 5. Prepare competency examples

    Prepare competency examples mapped to Marshall Wace's stated values

    For the 30-minute Human Capital Zoom call, prepare stories showing how you develop and execute new initiatives, voice differentiated opinions, collaborate in a team, and act with integrity.

    From .

Rehearse Marshall Wace quant math, Codility prep, and project defense out loud

Soreno's AI interviewer tests your probability, linear algebra, statistical modeling, and research project defense out loud. Soreno is not affiliated with Marshall Wace.

Questions people ask

Keep going