SIG Quant Research Intern Interviews: 60 Experiences

Shared by janebeingjane ·
Reports
60 interview reports
Level
Internship
Stages
OA → Phone screen → Onsite

SIG PhD quant research intern interviews since late 2023: an online test, then a short recruiter call with probability questions that cuts many.

At a glance

  • Process: an online test (17 questions in 60 minutes from mid-2024), then a recruiter or HR call of about 30 minutes, then phone rounds, sometimes a data round, and a final. Five PhD candidates were cut right at the recruiter call (04, 08, 09, 13, 14).
  • Asked recently: probability in 16 of the 20 current PhD reports, dice questions in 7, coin flips and random times or circle questions in 5 each, and hobbies, poker or games in 9. In 2025 three candidates got the same pair: the median of three uniforms and a five-dice reroll (04, 05, 07).
  • Difficulty: in three of the five recruiter-call cuts, the candidate felt every answer was right (09, 13, 14). One interviewer said ranking depends mostly on how fast you answer (59).
  • Practise first: which coin pattern shows up first, catching the bus to the movie and nine-sided dice: reroll toward a straight or a full house.

Coverage: 60 SIG quant research intern and related reports in one numbered sequence, each candidate counted once:

  • PhD interns, Nov 2023 to 2026 (01–20): current, and the base for the counts and tables below.
  • Master's and bachelor's interns (21–28), interns who don't state a degree (29–34, including two who posted between rounds) and Quant Researcher reports with no hiring level (35–40): each in its own section.
  • History, 2016 to Mar 2023 (41–60): PhD intern reports, outside the current counts.

PhD students who applied for full-time roles are in SIG PhD Quantitative Researcher New Grad OA / Interviews and SIG PhD Quantitative Researcher Experienced Hire OA / Interviews.

What came up

These counts cover the 20 current PhD reports (01–20). Probability dominates, and the recruiter call is the first filter.

Topic Reports Last seen
Probability 16 of 20 Nov 2026
Recruiter or HR call 12 of 20 Nov 2026
Online test 9 of 20 Nov 2026
Hobbies, poker and games 9 of 20 Sep 2025
Dice questions 7 of 20 Nov 2026
Statistics and estimation 6 of 20 Nov 2026
Coin-flip questions 5 of 20 Jan 2025
Random times, or points and lines on a circle 5 of 20 Jan 2025
Data exercise or data round 3 of 20 Nov 2026
Betting games, arbitrage, fair prices and market making 3 of 20 Jan 2025
Logic puzzles and brain teasers 3 of 20 Jan 2025
Motivation and fit (why SIG, why quant, start date) 3 of 20 Nov 2023
Python, programming and algorithms 2 of 20 Nov 2023

Rows: probability 01, 02, 04, 05, 06, 07, 08, 09, 10, 11, 13, 14, 15, 16, 19, 20; recruiter or hr call 01, 03, 04, 08, 09, 11, 13, 14, 16, 17, 18, 20; online test 01, 02, 03, 06, 07, 09, 12, 16, 20; hobbies, poker and games 05, 06, 07, 09, 13, 14, 16, 18, 19; dice questions 01, 02, 04, 05, 07, 13, 14; statistics and estimation 01, 05, 06, 07, 15, 20; coin-flip questions 08, 10, 11, 16, 19; random times, or points and lines on a circle 08, 09, 10, 11, 19; data exercise or data round 01, 07, 18; betting games, arbitrage, fair prices and market making 07, 09, 16; logic puzzles and brain teasers 07, 15, 16; motivation and fit 17, 19, 20; python, programming and algorithms 16, 20.

The recruiter or HR call

  1. Where it sits. In most reports from 2021 on, it is the first live step after the online test. A second recruiter or HR call, mostly on behavior and fit, can come a round or two later (16, 17, 20).
  2. What it holds. A few fit questions (hobbies, whether you play poker, why SIG), then two or three probability questions, often ones already shared online (01, 04, 08, 09, 11, 13, 14, 16, 18, 41, 42, 43, 45). A few HR calls had no technical questions at all (20, 44).
  3. How long. About 30 minutes (08, 09, 17); one took about 20 (14).
  4. It filters. Five candidates were cut right after it (04, 08, 09, 13, 14). In three of these, the candidate felt every answer was right (09, 13, 14).
  5. The 2019 exception. The STEM PhD internship candidate had no screening call. A campus recruiter's email led straight to phone interviews (51).

The Sydney report in "Not counted above" (29) also describes a 45-minute recruiter call on probability, but that candidate doesn't state a degree.

Questions we found

These tables cover the 20 current PhD reports.

Probability and counting

Question Reports Last seen
Coin flips: which pattern comes first (HHT vs HTT, HHT vs HTH, or a three-flip race) 4: 08, 10, 11, 19 Jan 2025
A bus to a movie, or two people arriving at random times 4: 09, 10, 11, 19 Dec 2024
Median of three uniforms in an interval 3: 04, 05, 07 Oct 2025
Five 9-sided dice: which one to reroll 3: 04, 05, 07 Oct 2025
Three random points on a circle: an obtuse triangle, or a triangle that holds the center 3: 08, 10, 11 Jan 2025
People around a circle pass a card left or right: how many end with none, or how many get one 2: 09, 19 Dec 2024
Two trips last year, at least one was international and in December: chance both were international (23/47) 2: 13, 14 Feb 2024
Two players take turns rolling a die and the first 6 wins: chance the first player wins; given that player wins, the expected number of rolls (61/11) and the chance it ended on roll one (11/36) 2: 13, 14 Feb 2024
10 dice: chance the sum is a multiple of 7, or of 6 1: 01 Nov 2026
X uniform on [0,1], Y uniform on [0,2]: chance X + Y is in [1,2] 1: 08 Jan 2025
Pick 3 of the numbers 1 to 20: chance one is the average of the other two 1: 09 Dec 2024

Statistics, linear algebra and numerics

Question Reports Last seen
Linear regression (no details shared) 1: 01 Nov 2026
A protractor reads a triangle as 56, 55 and 57 degrees: true angles? 1: 06 May 2025
Given X + Y for independent mean-0 normals, estimate X 1: 06 May 2025
Three variables with corr(X, Y) = corr(Y, Z) = 0.9: what is the lowest corr(X, Z) can be (0.62)? Is 0.5 possible? How do you check a matrix is a correlation matrix? 1: 16 Nov 2023
How long does computing a determinant take? 1: 16 Nov 2023

Logic puzzles

Question Reports Last seen
Is a three-person vote fair? 1: 01 Nov 2026
A, B and C each say something, not all true 1: 02 Oct 2026
A sudoku-like 4 x 4 grid with edge clues 1: 02 Oct 2026

Behavioral (BQ)

Question Reports Last seen
Hobbies, do you play poker, and do you play card or competitive games? 6: 05, 07, 09, 13, 14, 19 Sep 2025
Why SIG? Why quant research? Why trading? 3: 17, 19, 20 Nov 2023
When do you graduate, and when could you start? 1: 16 Nov 2023
Which office, and what kind of work do you want? 1: 17 Nov 2023
Would you rather see your strategy's impact up close or from further away? (from a reply) 1: 18 Nov 2023
Which other companies are you talking to, and how far along? 1: 20 Nov 2023

Betting games, arbitrage and fair prices

Question Reports Last seen
The AKQ game 1: 06 May 2025
Play a round of Acey-Deucy, a card betting game 1: 06 May 2025
2 red and 2 black cards turned over one by one, one bet that the next card is red: when do you bet? 1: 16 Nov 2023

Programming and algorithms

Question Reports Last seen
Median of a large data set, fast 1: 01 Nov 2026
Write your own range() in Python 1: 16 Nov 2023
pandas usage, matrix and vector sums, a sliding-window bonus question, as one-liners 1: 20 Nov 2023

Data exercise

Question Reports Last seen
A data project, then two data interviews on data handling and strategy 1: 01 Nov 2026
Discuss a paper dataset (expected value, risk) with no computer 1: 06 May 2025

Trading and business

Question Reports Last seen
How does SIG make money? Then a strategy of buying and selling calls built on that 2: 16, 20 Nov 2023
A round described as testing how you work out strategies in trading scenarios or games (announced for the next round) 1: 18 Nov 2023

20 PhD intern reports, Nov 2023 to 2026

01: Nov 3, 2026 · PhD QR intern · out at the data interviews

Steps: OA, HR call, phone screen, HR call, data project, two data interviews, onsite.

  1. First HR call: two or three easy probability questions.
  2. Phone screen (1 hour): median of a large data set, fast; 10 dice that win only on a multiple of 7; linear regression; is a three-person vote fair?
  3. Data interviews: friendly on data handling; on strategy, the interviewer misread an answer and kept pressing. No offer.

02: Oct 16, 2026 · PhD QR intern, 2027 cycle · passed the OA

OA (17 questions, 1 hour): mostly probability; the hard ones were conditional chances and dice or spinner expected values. Also A, B and C each say something, not all true, and a slow sudoku-like 4 x 4 grid.


03: Oct 13, 2025 · PhD intern, QR then ML program · offer

OA Jul 28; turned down for QR as not in the second-to-last PhD year, then invited to SIG's new ML program. HR call Sep 17, technical Sep 30, onsite Oct 9, offer Oct 13. The rest was hidden.


04: Oct 8, 2025 · PhD QR intern · cut at the recruiter screen

Recruiter screen (late September):

  1. Three uniforms on [0,3]: chance the median is in [1,2].
  2. Reroll one of five 9-sided dice showing 2, 2, 4, 6, 7; three of a kind plus a pair, or three in a row plus a pair, wins. Follow-ups: compare 2, 2, 2, 4, 6 with 2, 4, 4, 4, 6, then which new roll tips it.

Rejected two days later.


05: Sep 22, 2025 · PhD QR intern · first phone round

It opened with hobbies; the candidate felt card players have an edge. Then the median of three uniforms and the five-dice reroll, which takes a quick estimate. A reply: same two questions.


06: May 11, 2025 · PhD QR intern · no offer after the onsite

Two months, 11 rounds counted from the OA. Onsite in Philadelphia:

  1. Round 1: chat, then a protractor reads a triangle as 56, 55 and 57 degrees: true angles? You pick the assumptions.
  2. Round 2: chat, then given X + Y for independent mean-0 normals, estimate X, and the AKQ game.
  3. Round 3 (75 min): discuss a paper dataset (expected value, risk) with no computer, then play Acey-Deucy.

07: Jan 5, 2025 · PhD QR / systematic trading intern · out after the phone round

It opened with a few personal details and hobbies; the candidate noted SIG likes games, so poker or bridge can come up.

  1. Three uniform numbers on [0,1]: chance the median is in [a, b].
  2. Five 9-sided dice showing 3, 3, 5, 7, 8. Roll one die again each turn; three of a kind plus a pair wins, and so do three in a row plus a separate pair. Best strategy, and the expected number of rolls? Follow-up, without working it out: is 3, 3, 3, 5, 7 or 3, 5, 5, 5, 7 easier to win from?

The author later made the winning rule precise: the run of three and the pair must use different dice, so 3, 5, 5, 6, 7 does not win.

The candidate answered the follow-up by instinct, picking the hand with the 5s in the middle, then lost the thread after the interviewer went quiet. They felt the interviewer cared most about mathematical intuition. They also said SIG barely separates quant research from systematic trading, and described the usual process, as they understood it, as about eight rounds: the online test, two phone rounds on probability, statistics and brain teasers, a presentation of a data exercise, two or three strategy and market-making rounds, and possibly a team game session.


08: Jan 3, 2025 · PhD QR intern · cut at the first HR phone round

30 minutes, three standard questions:

  1. Flip a coin until HHT or HTT appears: chance HHT comes first.
  2. Three random points on a circle: chance they form an obtuse triangle, which is the chance they fit in one half of the circle.
  3. X uniform on [0,1], Y uniform on [0,2]: chance X + Y is in [1,2].

The candidate had seen similar questions before. Their hints: for the coins, condition on the first flip or set up a Markov chain; for the triangle, think of the half circle; for the sum, use the density of X + Y. They had heard SIG's process is long and gets harder as it goes.


09: Dec 6, 2024 · PhD quant intern, 2025 cycle · cut at the HR screen

30 minutes: fit questions such as hobbies, then three questions. HR said every answer was right; a rejection came two days later.

  1. Pick 3 of the numbers 1 to 20: chance one is the average of the other two.
  2. Bob and a bus to a 7 pm movie: Bob reaches the stop between 6 and 7, the bus comes between 6 and 7 and waits 10 minutes, and the ride takes 20.
  3. 100 people in a circle each pass a card left or right: expected number left with none.

A reply from another candidate: the same quick answers and the same rejection two days later. A third candidate replied that the year before they reached the sixth round, while this year they were turned down right after the online test.


10: Aug 29, 2024 · PhD QR intern, 2025 cycle · past the test, then a phone screen

The test had 17 probability questions. They think they missed about two and still moved on. The phone screen had three classic questions: two people meeting at random times, HHT before HTH and whether three random points on a circle make a triangle that holds the center.


11: Mar 11, 2024 · PhD QR intern, through a headhunter · passed the HR math round

The HR math round had three questions: a bus with an arrive-by deadline, the triangle holding the center of a circle and HTH before HHT. They passed two days later. Another candidate with the same round gave the full order: an HR math round, a recruiter behavioral round, a quant team round, then a final quant round.


12: Mar 3, 2024 · PhD QR intern · online test

They took the test and asked others for their results. No questions were shared.


13: Feb 12, 2024 · PhD QR intern · out after the HR round

The HR round asked about hobbies and whether they play poker, then two probability questions and a follow-up: the two-trips question (23/47) and the die race, where the first 6 wins: given the first player wins, the expected number of rolls (61/11), then the chance it ended on the first roll (11/36). They answered everything right away and were still turned down the next morning.


14: Feb 3, 2024 · PhD QR intern · out after the recruiter round

The test had one section of 9 questions, the same as ones already shared but with new numbers. A week later came a recruiter round; the email said 2 to 3 math questions in 30 minutes. It opened with hobbies and poker (they said they don't play), then the two-trips question, the chance the first player wins the die race, and the expected rolls given that player wins. The recruiter agreed with every answer and the call took about 20 minutes. A rejection came two working days later. A reader replied that you should say you like card games, because SIG's culture is built around them.


15: Dec 5, 2023 · PhD QR intern · third round

A phone round that walked through the resume, then two to four questions across math, statistics, probability and logic, then time for the candidate's questions. No question text was shared.


16: Nov 30, 2023 · PhD QR intern, a PhD-hire internship in the Bala Cynwyd office · five rounds, then out

All five rounds were by phone.

  1. Online test.
  2. Recruiter call: HR questions and two probability questions like ones already shared.
  3. Recruiter behavioral round. It asked their graduation date and seemed surprised the application was for an internship.
  4. A quant: 2 red and 2 black cards turned over one by one with one chance to bet the next is red; a brain teaser; a coin question. It used 30 of the 60 minutes, and the next round was described as similar.
  5. A senior quant: what they knew about SIG and market making, writing their own range() in Python, a calculus-heavy probability question they finished after hints in about 20 minutes, and how to check that a matrix is a correlation matrix (positive semidefinite), then how long a determinant takes to compute.

A rejection came a week later. A peer in the same field was moved to the trader track after round 3 and played a market-making game in round 4.

Practice: What do you know about SIG.


17: Nov 28, 2023 · PhD QR intern · passed the recruiter behavioral round

The second call was a 30-minute recruiter behavioral round that went down a checklist: which office, what kind of work, why quant research, why SIG. They heard they'd passed within an hour and were told about four rounds were left.

Practice: Why SIG?, Why quant research?.


18: Nov 22, 2023 · PhD QR intern · recruiter call, then a games round

After the test, a recruiter call: 30 minutes of behavioral questions, then two problems they hadn't seen, solved with hints. The recruiter helped them fix mistakes. They were told three rounds were left: one more technical round, then a data exercise. The email for the next round said behavioral plus technical, with the technical part testing how well you can work out strategies in trading scenarios or games. Other candidates added:

  • In the behavioral part, one was asked whether they'd rather see their strategy's impact up close or from further away, and was told SIG considers quant research and quant systematic trading at the same time.
  • The data exercise is followed by four interviews over two days.

Practice: Would you rather see your strategy's impact up close or from further away?.


19: Nov 17, 2023 · PhD QR intern · passed the first phone round

Booked for 20 to 30 minutes, it ran past 40. It started with hobbies, why SIG and why quant research, and the interviewer kept coming back to whether they like poker. Then three questions: a three-flip coin pattern race (details forgotten), waiting for a bus (set up the limits carefully), and people around a circle passing a card, asking how many people get one. They passed the next day.

Practice: Coin flips: which pattern comes first (HHT vs HTT, HHT vs HTH, or a three-flip race), Why SIG?, Why quant research?.


20: Nov 6, 2023 · PhD QR intern · five rounds, then out

  1. Online test, result within a week.
  2. Questions already shared online, three days later.
  3. HR: background, other interviews, why SIG, projects. No technical questions.
  4. Coding: programming background, how they use pandas, adding matrices and vectors, and a bonus sliding-window question. The interviewer pushed for short, one-line answers.
  5. HR again: behavioral questions, then how does SIG make money, then analyzing a strategy of buying and selling calls built on that answer.

They were out within a week. A take-home assessment would have been next. Another candidate was told their fourth round would be 2 to 4 questions on linear algebra, probability, calculus and statistics.

Practice: How does SIG make money?, Why SIG?, Why quant research?.


Master's and bachelor's interns

Eight more SIG quant research intern reports from candidates who say they were master's students (seven) or a bachelor's student (one, 22). Dates are report dates.

Online test puzzles (unlinked)

Question Reports Last seen
Legs in a barn or cage (520 legs): how many of one animal? 1: 24 Sep 2024
A balanced hanging mobile (96 lb): weight of one shape 1: 24 Sep 2024
A canoe or rowing trip: work out the start time 1: 25 Sep 2024

Probability and counting

Question Reports Last seen
A bus to a 7 pm movie, or reaching a cinema by bus before the film starts 2: 21, 22 Jan 2026
A spinner with chances 1/4, 1/4 and 1/2, or 1/5, 2/5 and 2/5 2: 24, 26 Sep 2024
Bold betting from 3 tokens to 5 2: 25, 26 Sep 2024
Two flowers that open at random times: chance both are open together 2: 24, 25 Sep 2024
Expected pairs in a hand dealt from a deck of pairs 2: 24, 25 Sep 2024
Two dice summing to a given number 1: 21 Jan 2026
Roll three six-sided dice: chance one is the average of the other two 1: 22 Jun 2025
The higher of two dice 1: 24 Sep 2024
Paths to (7, 4) with at most two steps in a row one way 1: 24 Sep 2024
Enough muffins: 60% croissants, 2 muffins, 5 people 1: 25 Sep 2024
Two factories making red items: which factory did they come from? 1: 24 Sep 2024
An integral for the expected chord length (practice) 1: 27 Mar 2024
A conditional probability about a painted cube 1: 27 Mar 2024

Puzzles and arithmetic (online test)

Question Reports Last seen
A boat or canoe going up and down a river 2: 24, 25 Sep 2024
Legs in a barn: 520 legs of cows, chickens and spiders 1: 24 Sep 2024
Five people at a table: who sits next to whom 1: 24 Sep 2024
A hanging mobile weighing 96 lb 1: 24 Sep 2024
A grid of buildings with 1 to 4 floors 1: 26 Sep 2024

Logic and trading

Question Reports Last seen
Is a three-person vote fair? 1: 22 Jun 2025
The fair price of a call option 1: 27 Mar 2024

21: Jan 15, 2026 · master's QR/QST intern · cut at the HR screen

The candidate said the OA questions had been shared online before. HR asked the chance of reaching a cinema by bus before the film starts, and the chance two dice sum to a given number.


22: Jun 14, 2025 · bachelor's systematic trading / QR intern · out after three rounds

Summer 2025 role; two video rounds and an onsite would have come next. Questions across the rounds (the post doesn't say which round asked which):

  1. Roll three six-sided dice: chance one is the average of the other two.
  2. Bob and a bus to a 7 pm movie: the bus comes at a random time between 6 and 7, always waits 10 minutes, and the ride takes 20.
  3. Is a three-person vote fair? Every voter ranks A, B and C; first choices tie, and so do second choices. A proposes B against C first, then the winner against A.

The third round also had an algorithm question, answered in words only. The candidate thinks they missed at most one online test question.


23: Oct 4, 2024 · master's QR intern · reached the final

The author had reached SIG's final round for a quant research internship and asked what it would be like. On an earlier round they said it was behavioral questions first, then pure probability.

Two other candidates added their own steps in the replies:

  1. One had passed the online test and a first technical phone round: 45 minutes, three probability questions, similar to ones shared online. Their next round was a chat with someone from the recruiting team about background and career interests.
  2. Another described that recruiting-team chat: the recruiter goes over your past projects, suggests a few teams that fit, and asks for one or two preferences; team matching comes only after all the interviews. This person was cut after their first round, so this may be secondhand.

24: Sep 14, 2024 · master's QR intern, 2025 cycle · online test

A 60-minute test with 17 questions, most already shared online. The candidate listed what they remembered: legs in a barn (520 legs of cows, chickens and spiders), five people at a table, a hanging mobile weighing 96 lb, a simple conditional probability, two factories making red items, the higher of two dice, a boat going up and down a river, a spinner with chances 1/4, 1/4 and 1/2, two flowers that open within 20 hours and stay open 4 and 8 hours, paths to (7, 4) with at most two steps in a row one way and pairs in 5 cards drawn from 5 pairs.


25: Sep 14, 2024 · master's QR intern · online test

The same 17-question, 60-minute test. They called it tight on time and wrote down the slowest questions: enough muffins (60% croissants, 2 muffins, 5 people), a canoe trip with a 2 mph current, two flowers in bloom over 30 days, bold betting from 3 tokens to 5 and pairs in 4 cards from a 10-card deck.


26: Sep 10, 2024 · master's QR intern · online test, not passed

17 questions in 60 minutes, some new. They shared three: a spinner with chances 1/5, 2/5 and 2/5, a grid of buildings with 1 to 4 floors, and bold betting to 5 tokens.


27: Mar 19, 2024 · master's QR intern · passed the recruiter round

A 9-question, 60-minute test, then a recruiter round under an hour with three questions: an integral for the expected chord length, a conditional probability about a painted cube, and the fair price of a call option. The recruiter gave hints when they got stuck.

Practice: practice.


28: Feb 1, 2024 · master's QR intern · online test

Only one detail: the test lets you mark questions to revisit, and the summary page counts attempted and revisited questions.


Interns who don't state a degree

These four SIG quant research intern reports never say whether the candidate was a PhD student. Dates are report dates unless marked.

Two of these, 31 and 34, posted while waiting for a later round and share what SIG told them to expect.

Other

Question Reports Last seen
Expected length of circle arcs colored between random points and their nearest neighbors 1: 32 Aug 2026

29: Sydney, a recruiter call on probability between the test and the technical round

Report date: Oct 30, 2026 · Role: Quantitative Researcher Intern · Location: Sydney · Degree: not stated · Result: No offer

Online assessment (1 hour)

  1. 17 questions on logic, combinatorics and probability.

Recruiter interview (about 45 minutes)

  1. The call came after the online test and before the technical interview.
  2. Probability questions asked by the recruiter. No specific questions were shared, and the review doesn't mention behavioral ones.

Technical interview (1 hour)

  1. Probability, statistics, linear algebra and algorithms.
  2. The candidate found the interviewers kind and supportive.

The process ended after this round. The candidate rated the experience positive and the difficulty average.


30: A 17-question test where a few questions eat the clock

Report date: Sep 17, 2026 · Role: Quantitative Researcher Intern · Degree: not stated · Result: No offer after the OA

Online assessment (60 minutes)

  1. 17 probability and logic questions.
  2. Most felt easy to medium. Three or four took much longer than the rest.
  3. Some were easiest to solve with a recursion, meaning you write the answer in terms of a smaller version of the same problem.

The candidate didn't reach a recruiter call.


31: Dublin, Summer 2027, a trader round next

Report date: Aug 22, 2026 · Role: Quantitative Research / Quantitative Systematic Trading intern, Summer 2027 · Location: Dublin · Result: Not reported

  1. Invited to a second-round technical interview with a trader.
  2. The recruiter said a quant research interview and a data-science task would come after that.

The post doesn't describe the earlier rounds.


32: London, a quiz with PDEs and a hard circle question

Interview date: Aug 2026 · Report date: Sep 14, 2026 · Role: Summer Quantitative Research Intern · Location: London · Degree: not stated · Result: No offer

Online quiz

  1. Questions on economics, partial differential equations (PDEs), algorithms and probability.

Zoom interview

  1. Maths, statistics and algorithms questions.

The one question shared (round not stated)

  1. Put n random points on a circle. Each point colors the arc between itself and its nearest neighbor. What is the expected total length of the colored part of the circle? The candidate called it difficult. Practice: expected colored arc length between random points on a circle.

The whole process took under a month. No recruiter or HR call is described.


33: New York, one day from test to result

Interview date: Aug 2026 · Report date: Aug 20, 2026 · Role: Quantitative Researcher Intern · Location: New York · Degree: not stated · Result: No offer after the OA

Online assessment (Mettl, an online test platform, 60 minutes)

  1. 17 questions.
  2. Counting, Bayesian and basic probability (Bayesian here means updating a chance after you learn something new).
  3. Basic arithmetic and basic calculus.

The candidate applied through a source they didn't name, and the whole process took one day. They felt the test has been getting harder than the older SIG questions you can find online.


34: Summer 2026, a 45 to 60 minute second round

Report date: Feb 18, 2026 · Role: Quant Research Intern, Summer 2026 · Location: Not given · Result: Not reported

  1. A second-round phone interview was scheduled for 45 to 60 minutes.
  2. The candidate described it as a round on probability, statistics and logic.

The post doesn't describe the earlier rounds or how this one went.


Quant Researcher reports with no hiring level

These six reports are for Quant Researcher roles (three of them for QR or Quant Systematic Trader) and never say whether the job was an internship or full-time. One candidate was on a PhD track (36). Three were posted before the next round and only say what the candidate was told to expect.

Probability and counting

Question Reports Last seen
Count up-and-right grid paths from (0,0) to (6,4) that avoid a blocked middle point 1: 35 Aug 2026
A complicated dice game asked in a recruiter phone screen (rules not shared) 1: 39 Aug 2025

Trading discussion

Question Reports Last seen
Where trading strategies come from, which would make more money, and which would swing more in value 1: 37 Dec 2025

35: London, a timed test with a blocked-grid question

Interview date: Aug 2026 · Report date: Sep 9, 2026 · Role: Quant Researcher (hiring level not reported) · Location: London · Result: Did not move past the OA

Online assessment (timed)

  1. Probability, mathematical reasoning and number questions under a tight time limit.
  2. A grid question the candidate called a probability question: paths from (0,0) to (6,4) where one point in the middle can't be used. The report doesn't say which point or exactly how the path may move. Practice: count grid paths that avoid a blocked point.

The candidate applied online through SIG's careers site.


36: PhD QR/QST candidate, a quant call after the recruiter round

Report date: Jul 30, 2026 · Role: Quantitative Research / Quantitative Systematic Trading, PhD track (hiring level not reported) · Location: Not given · Result: Not reported

  1. Recruiter round: passed. The candidate didn't share what it covered.
  2. Next, a phone interview with a quant: described to the candidate as two to four questions on probability, statistics and logic.

37: Philadelphia, OA then HR then a researcher

Interview date: Dec 2025 · Report date: Jun 13, 2026 · Role: Quant Researcher (hiring level not reported; the candidate was a student) · Location: Philadelphia · Result: No offer

Online assessment

  1. Passed. No details given.

HR interview

  1. Named only as an HR interview. No content was shared.

Interview with a researcher

  1. General questions about trading strategies: where they come from, which would make more money, and which would swing more in value (be more volatile). The report lists these without naming the round.

The candidate came through on-campus recruiting, and the process took one to two months.


38: QR/QST, invited to a 45-minute first recruiter call

Report date: Dec 3, 2025 · Role: QR / QST (hiring level not reported) · Location: Not given · Result: Not reported

  1. Invited to a first-round recruiter call of about 45 minutes.

The candidate didn't post how it went. The replies suggested preparing for behavioral questions, the CV and easy probability, but none of them said they had been asked those things.


39: A recruiter phone screen with a dice-game question

Interview date: not given (posted right after the call) · Report date: Aug 5, 2025 · Role: Quant Researcher / Quant Systematic Trader (hiring level not reported) · Location: not given · Result: not reported

Recruiter phone screen

  1. A behavioral part. The candidate felt it went well but didn't share the prompts.
  2. A first question the candidate answered right away. It isn't described.
  3. A second question about a complicated dice game. The candidate got stuck but talked the recruiter through their reasoning. The rules and numbers weren't shared.

The post doesn't say how long the call ran or what order these parts came in. The recruiter said to expect an answer in about a week, maybe sooner. In a reply, another candidate said they also got stuck on a question in this call and still made it to the final round.


40: QR, told what the quant interview would cover

Report date: Nov 19, 2024 · Role: Quant Researcher (hiring level not reported) · Location: Not given · Result: Not reported

  1. HR interview: done. The candidate didn't share what it covered.
  2. Next, an interview with a quant: HR said it would cover logic, linear algebra, probability and statistics.

History (2016–Mar 2023)

These 20 PhD intern reports are more than three years old, so they stay out of the counts above. Their named questions:

Online test puzzles (unlinked)

Question Reports Last seen
Three dice paying $20 for a triple and $10 for a pair: expected return 2: 49, 54 Nov 2020

Probability and counting: asked more than once

Question Reports Last seen
Two envelopes, one holds twice the other, the smaller amount is uniform on 0 to 100: when should you switch? 2: 43, 46 Mar 2022
Estimate π, or e, with a Monte Carlo simulation 2: 45, 59 Nov 2021
Coin flips: which pattern comes first (HHT vs HTT, HHT vs HTH, or a three-flip race) 2: 48, 52 Dec 2020
10 dice: chance the sum is a multiple of 7, or of 6 2: 53, 55 Apr 2019
Random chords in a circle: expected number of regions (three chords give 5; six random points cut a pie into how many pieces?) 2: 53, 56 Apr 2019
400 coin flips: chance of at least 220 heads 2: 54, 59 Mar 2019

Probability and counting: asked once

Question Reports Last seen
Three random points on a circle: an obtuse triangle, or a triangle that holds the center 1: 41 Mar 2023
Chance at least two of three independent variables meet a condition; E[max(X, Y)] for two correlated normal variables 1: 45 Nov 2021
Conditional probability: at least one daughter, versus a daughter opened the door (online test) 1: 45 Nov 2021
An urn with 2 red and 1 blue ball, each drawn ball is replaced by a blue one: expected draws until no red is left 1: 46 Oct 2021
16 flips of a biased coin: expected number of runs of heads 1: 46 Oct 2021
5 of 30 people picked at random happen to be in height order: chance the first one is the shortest of all 30 1: 46 Oct 2021
Expected distance between two random points on a circle (4/π for a unit circle) 1: 47 Sep 2021
A draws three different digits from 1 to 9, B from 1 to 8, each makes the largest 3-digit number: chance A wins 1: 49 Nov 2020
A coin whose two faces are each heads or tails at random: the first flip is heads, chance the second is heads (3/4) 1: 50 Oct 2020
2 red and 2 blue balls drawn one at a time, you guess each color: expected right guesses (17/6) 1: 54 Mar 2019
Expected rolls to see a 6, and to see two 6s in a row 1: 54 Mar 2019
A walker at a hub O moves to A, B or C at random and always comes back: expected time from A to C (6) 1: 54 Mar 2019
You win each game with chance 60%: best of three, or play until someone leads by two? 1: 54 Mar 2019
100 teams: the fewest games to find the best and the second best team 1: 58 Dec 2018

Statistics, linear algebra and numerics

Question Reports Last seen
Three variables with corr(X, Y) = corr(Y, Z) = 0.9: what is the lowest corr(X, Z) can be (0.62)? Is 0.5 possible? How do you check a matrix is a correlation matrix? 2: 48, 54 Dec 2020
Estimate London's population, explaining each step 1: 42 Apr 2022
A matrix A with standard normal entries: dimension of the matrices B with AB = BA 1: 43 Mar 2022
How uncertain are regression estimates? Bootstrap 1: 43 Mar 2022
Rank, diagonalizability and Jordan form (online test) 1: 45 Nov 2021
Draw two correlated normal variables with given means, variances and covariance, without library functions 1: 47 Sep 2021
How do you handle missing data? 1: 48 Dec 2020
A square matrix with A² = 0: the most nonzero entries it can have 1: 53 Apr 2019
Estimate n from m numbers drawn from 1 to n (largest value or the mean), and how to judge an estimator 1: 58 Dec 2018
Does A independent of B and of C mean A is independent of B and C together? 1: 58 Dec 2018
A finite-difference derivative: the best step size (rounding error against truncation error) 1: 58 Dec 2018

Behavioral (BQ)

Question Reports Last seen
Why SIG? Why quant research? Why trading? 2: 58, 60 Dec 2018
Is your research theory or numerics? Summarize your papers; would you relocate; how was your last internship; which tools do you use most? 1: 44 Feb 2022
Hobbies, do you play poker, and do you play card or competitive games? 1: 47 Sep 2021
When do you graduate, and when could you start? 1: 47 Sep 2021
Which other companies are you talking to, and how far along? 1: 49 Nov 2020
Questions on interest in SIG and in the work itself (practice) 1: 51 Sep 2020
Your research, visa status, studies and plans 1: 56 Jan 2019
What do you value most in a company? 1: 58 Dec 2018

Betting games, arbitrage and fair prices

Question Reports Last seen
A stock at $100 will be $50 or $200 tomorrow, and a casino takes even-money bets on its direction: how do you lock in a profit? 2: 49, 54 Nov 2020
Roll 10 or 12 dice: you're paid $10 if the sum is low (at most 30, or under 30), otherwise you pay $1. Do you play? 2: 50, 53 Oct 2020
Name a card from a 52-card deck and win $1 when it comes up: a fair price 1: 47 Sep 2021
Two boxes of money: you and a friend both win only if you pick different boxes 1: 47 Sep 2021
Pay $2 to reroll your die, then the opponent rolls and the higher roll is paid the difference: when do you reroll? 1: 49 Nov 2020
A friend doubles the pot: the lowest chance of winning at which you should call (1/4) 1: 52 Jan 2020
A contract that pays like a call with a $22 strike on a stock with a few possible prices (0.8) 1: 52 Jan 2020
Three bonds with given prices and payoffs when the market goes up, stays flat or goes down: how do you build a portfolio? 1: 53 Apr 2019
Three horses pay 2, 4 and 6 times a $1 ticket: how do you bet? 1: 54 Mar 2019
A whole number X from 1 to 10: you win X if it is prime, otherwise you lose X/2 1: 55 Feb 2019
You and the interviewer each flip 3 coins: equal head counts pay the interviewer $2, otherwise you get $1 1: 59 Feb 2018
Each of two players has two dice; you lose $1 if any of your numbers matches one of mine. Do you play? 1: 60 Dec 2016
Two buttons for a $60 box and a $40 box, both players press at once without talking: what do you press? (The author later answered: press $60 with chance 3/5.) 1: 60 Dec 2016

Programming and algorithms

Question Reports Last seen
Python: list vs tuple, Python 2 vs 3, Python vs C++ 2: 53, 58 Apr 2019
Python details: list comprehension, range vs list, del, NumPy speed and memory layout, pandas inner vs outer join 1: 43 Mar 2022
The median of 16 MB of data with 4 MB of memory; sorting algorithms 1: 43 Mar 2022
What happens if you assign print = 5 in Python 3? 1: 58 Dec 2018

Data exercise

Question Reports Last seen
Simulated price data with traps: a stock split, missing data, correlated stocks, sampled bid and ask prices 1: 43 Mar 2022
A data set and a prompt, about a week of reading and coding 1: 51 Sep 2020
Stock data with something to predict, over one week 1: 57 Jan 2019

Trading and business

Question Reports Last seen
What do you know about SIG and market making? 2: 58, 60 Dec 2018
What do you know about options? 1: 53 Apr 2019

41: Mar 20, 2023 · PhD QR intern · first HR call

Small talk, then two or three probability questions in the style of the well-known green book, using total expectation and total probability. One was whether three random points on a circle make a triangle that holds the center.


42: Apr 27, 2022 · PhD QR intern · out after round 2

A 60-minute math test with questions already shared online. Round 1 was a recruiter video call; the camera was expected though nobody said so. It had behavioral questions and a simple Bayes question. Round 2, two days later, was with a lead: about 20 minutes on research, then a four-part estimation exercise with a market-making flavor (estimate London's population), explaining every step. A rejection came 8 days later with no feedback.


43: Mar 3, 2022 · PhD QR intern · data exercise, then out

This took two months.

  1. A one-hour online test in four sections.
  2. HR: a ball-drawing question and the two-envelope paradox (two envelopes with a known range).
  3. Programming background: sorting algorithms, the median of 16 MB of data using 4 MB of memory, and how uncertain regression estimates are (bootstrap).
  4. A senior quant: a green-book race probability, many Python questions (list comprehension, tuples, range vs list, del, why NumPy is fast and how it lays out memory, how much memory an object takes, pandas inner vs outer join), and a linear algebra question: for a matrix A with standard normal entries, the dimension of the matrices B with AB = BA.
  5. They were moved from quant research to the quant strategist track.
  6. Data exercise, from a Wednesday to the following Sunday: simulated price data with traps such as a stock split, missing data, correlated stocks and sampled bid and ask prices. They were told 10 to 15 hours. It took them 40 to 60.
  7. A one-hour discussion of the exercise with a quant.

A rejection came three days later. The onsite would have been four rounds: math, coding, and two rounds on trading and game theory.


44: Feb 17, 2022 · PhD QR intern · passed the first HR call

A 30-minute HR call with no technical questions: is your research theory or numerics, summarize your projects and papers, would you relocate, how was your past internship and would you go back, and which languages and tools do you use most. The recruiter then explained SIG's business and said four more interviews would follow.


45: Nov 8, 2021 · PhD QR intern, 2022 cycle · out after round 1

  1. A one-hour test in four sections. Two sections went fast with a computer algebra tool. Others had conditional probability (at least one daughter, versus a daughter opened the door) and linear algebra (rank, diagonalizability, Jordan form).
  2. HR: two questions already shared online, solved quickly. They were told on the spot that the next round would follow.
  3. Technical round: research first, then the chance at least two of three independent variables meet a condition, generating e with a Monte Carlo simulation, and E[max(X, Y)] for two correlated normal variables with given means and covariance.

They answered everything and were rejected four days later.


46: Oct 22, 2021 · PhD QR intern, 2022 cycle · first round

Their questions: 5 of 30 people picked at random happen to stand in height order, so what's the chance the first of the five is the shortest of all 30; and the expected number of runs of heads in 16 flips of a biased coin. They also passed on a teammate's questions: an urn with 2 red and 1 blue ball where each drawn ball is replaced by a blue one (expected draws until no red is left), and the two-envelope problem with the smaller amount uniform on 0 to 100.


47: Sep 8, 2021 · PhD QR intern, 2022 cycle · out at the first technical round

The test used questions that had been shared online for years. The behavioral round went down a list; they felt SIG likes competitive people: whether you play cards or online battle games, whether you want to win, how far you or your team got. The first technical round had four questions: the expected distance between two random points on a circle (4/π for a unit circle), drawing two correlated normal variables with given means, variances and covariance without library functions, a 52-card game where you name a card and win $1 when it comes up (a fair price), and two boxes of money where you and a friend both win only if you pick different boxes. They think saying they could only start the following year hurt.


48: Dec 3, 2020 · PhD QR intern, 2021 cycle · passed the phone round

Three questions: how to handle missing data, HHT before HTH, and the lowest possible corr(X, Z) when corr(X, Y) = corr(Y, Z) = 0.9 (the correlation matrix's determinant can't be negative, which gives 0.62). They moved on the next day.


49: Nov 26, 2020 · PhD QR intern, 2021 cycle · out after the technical phone round

The recruiter call asked directly which other companies and stages they were in. The technical round had three questions: the $100 stock that goes to $50 or $200 with an even-money casino bet (find the arbitrage); A draws three different digits from 1 to 9 and B from 1 to 8, each makes the largest 3-digit number, chance A wins; and a die game where you can pay $2 to reroll before the opponent rolls and the higher roll is paid the difference (when to reroll).


50: Oct 30, 2020 · PhD QR intern, 2021 cycle · phone round after HR

About 20 minutes on research; the interviewer wanted to know whether it created genuinely new tools. Then two questions: a coin whose two faces are each set to heads or tails at random, the first flip shows heads, chance the second does too; and roll 12 dice, win $10 if the sum is under 30, otherwise lose $1, do you play?


51: Sep 15, 2020 · PhD quant research intern, summer 2020 · offer, then a return offer

Interview date: Oct–Nov 2019 · Role: 10-week quant research internship for STEM PhD students (summer 2020) · Experience: fifth-year math PhD student · Location: U.S. (onsite at a SIG office)

Recruiter contact (not a call)

  1. After a department industry talk, the candidate met a SIG campus recruiter.
  2. The recruiter found the candidate's CV online, answered questions by email the same night, and invited them into interviews with no formal application. The candidate picked a start date two weeks later.
  3. The recruiter said to expect probability, statistics, calculus and linear algebra, possibly some computing, and no finance.

Nothing in this step was scored.

Three phone interviews (about one a week through October)

  1. Mostly probability and statistics, nothing very advanced: no measure theory and no stochastic differential equations.
  2. The candidate needed hints on several questions, and taking hints didn't rule them out.
  3. They also had to show real interest in the industry and in SIG, and explain their thinking clearly.

Take-home data exercise (about a week)

  1. After the third phone interview, SIG sent a data set and a prompt.
  2. The candidate spent several hours a day for about a week reading up on methods and writing code to analyze it. Python data tools such as pandas helped.

Onsite day (November 2019, about 5 hours)

  1. Five interviewers, plus informal talks with two more people.
  2. More math questions, several computing questions and some logic puzzles.
  3. Questions on why they were interested in SIG and in the work (What interests you about SIG and the work you would do here?).

The candidate said they couldn't share specific questions. They also said SIG treats graduate research interns as future full-time hires, so the intern process matched the full-time one, and that spots were filled on a rolling basis. A full-time PhD candidate from 2020 described a similar loop in the new grad roundup.


52: Jan 27, 2020 · PhD QR intern, summer 2020 · phone round 1

Three questions: when to accept when the pot is doubled (a win chance of at least 1/4), a contract that pays like a call with a $22 strike on a stock with a few possible prices (0.8), and HHT before HTT.


53: Apr 18, 2019 · PhD QR intern, summer 2019 · out after three phone rounds

SIG had dropped the outside professor screen by then: an HR call, then technical phones.

  1. A quant: the chance 10 dice add up to a multiple of 6, and how many pieces six random points on a circle cut a pie into.
  2. A second quant: research, Python vs C++, roll 10 dice and get $10 if the sum is at most 30 or pay $1 (do you play?), and three bonds with given prices and payoffs when the market goes up, stays flat or goes down (how do you build a portfolio?).
  3. A senior quant: research, programming basics, what they knew about options, a short company intro, then list vs tuple, Python 2 vs 3, and the most nonzero entries a square matrix with A² = 0 can have.

54: Mar 31, 2019 · PhD QR intern, fall 2018 cycle · passed the first technical round

A relaxed HR call about the process, then a first technical round run by an outside academic SIG used as a first filter. After background and research, and a note that SIG wants the smartest people:

  1. 400 coin flips: chance of more than 220 heads (about 2.5%).
  2. 2 red and 2 blue balls drawn one by one, guess each color: expected right guesses (17/6).
  3. Three horses paying 2, 4 and 6: how do you bet?
  4. Expected rolls to get a 6, and to get two 6s in a row.
  5. A stock at $100 goes to $50 or $200, and a casino takes bets on its direction: what do you do? (Buy the stock and bet $75 that it falls.)
  6. A hub O linked to A, B and C; each step takes one unit of time: expected time from A to C (6).
  7. You win each game with chance 60%: best of three, or play until someone leads by two?
  8. corr(X, Y) = corr(Y, Z) = 0.9 and corr(X, Z) = 0.5: is that possible?

55: Feb 4, 2019 · PhD QR intern · phone screen

Two questions: a whole number X uniform on 1 to 10, you win X if it's prime and lose X/2 otherwise, find the expected payoff; and the chance 10 dice add up to a multiple of 6.


56: Jan 26, 2019 · PhD QR intern · one-question phone round, then out

47 minutes. The first half hour was about research, visa status, studies and plans. Then one question: one line cuts a circle into 2 regions, so how many regions do three random lines through it make on average? They answered 5 and the interviewer agreed. HR said they'd send next steps; a rejection came two days later.


57: Jan 2, 2019 · PhD QR intern, by referral · data exercise

The data exercise was stock data with something to predict. They were given a week and could spend as much of it as they liked. A later round discussed the exercise and then asked programming and probability questions.


58: Dec 21, 2018 · PhD QR intern, by referral · passed the last phone round

The last phone round was with a senior quant. It started with what they knew about SIG, why quant, and what they value most in a company; there were no research questions. Then programming: they said Python and C++, and heard SIG uses Python a lot and some MATLAB. Python 2 vs 3, list vs tuple, and what happens if you assign print = 5 in Python 3. Three math questions:

  1. 100 teams: the fewest games to find the best and the second best.
  2. m numbers drawn from 1 to n: how do you estimate n? (The largest value, or twice the mean.) Then how to judge an estimator: bias, variance, consistency. Asked later whether a number could be drawn twice, the author said they took it as a simple random sample, so repeats were allowed.
  3. A is independent of B and of C. Is A independent of B and C together? Not always; they gave a two-dice example.

In a reply on another candidate's post, the same author said an earlier SIG phone round, about a week before this one, asked for the best step size of a finite-difference derivative. They framed it as trading rounding error against truncation error, which comes down to minimizing an expression like a·h² + b/h.

Practice: What do you know about SIG, Why SIG?, Why quant research?, What do you value most in a company?.


59: Feb 22, 2018 · PhD QR intern · out at the second phone round

Three questions: 400 coin flips, chance of at least 220 heads; estimate π with a Monte Carlo simulation; and you and the interviewer each flip 3 coins, chance you get the same number of heads. They knew how to do the second one but described it too loosely at first and lost time. The interviewer said the ranking depends mostly on how fast you answer; faster candidates get a fourth and fifth question.


60: Dec 20, 2016 · PhD QR intern · 40-minute phone screen

About 15 minutes on the resume: a machine learning internship project, why trading. Then two questions with pen and paper: each player rolls two dice and you lose $1 if any of your numbers matches one of mine (do you play?), and two buttons for a $60 box and a $40 box, both players press at once without talking (if both press the same button that box opens and you split it; if you press different buttons both boxes open and you split $100). It ended with a short talk about market making. Asked later how they answered the button question, the author said it is a mixed strategy: press the $60 button with chance 3/5 and the $40 button with chance 2/5.

Practice: What do you know about SIG, Why SIG?, Why quant research?.


Patterns across the PhD reports

  1. The online test changed shape over the years. In 2021 and 2022 it was one hour in four sections (43, 45), with questions that had been shared online for years (42, 47). In early 2024 one candidate had a single section of 9 questions (14). From mid-2024 it was 17 questions in 60 minutes (10, 02). The 2019 STEM PhD intern (51) had no online test at all and went from a recruiter email to phone interviews.

  2. The same classic questions keep coming back. The coin pattern race shows up in six reports from 2020 to 2025. The bus question and the three-points-on-a-circle question show up in four each. In 2025, three candidates got the same pair: the median of three uniforms and the five-dice reroll (04, 05, 07).

  3. A long loop for those who get past the first calls. Candidates who went further describe a data exercise and more interviews after it (01, 06, 43). Others went through five phone rounds before being cut (16, 20). The data exercise took about a week (51, 57); one candidate was told 10 to 15 hours and spent 40 to 60 (43).

  4. Speed and clear thinking count, and hints are normal. One interviewer said the ranking depends mostly on how fast you answer (59). Taking hints did not rule a candidate out (18, 51).

Common questions

How long is the SIG quant research intern online test?

From mid-2024 it was 17 questions in 60 minutes (02, 10, and 29, 30, 33 among interns who don't state a degree); a 2024 test had 9 questions (14).

What does the SIG recruiter call ask interns?

A few fit questions, such as hobbies, poker or why SIG, then two or three probability questions; it lasted about 30 minutes, and five current PhD candidates were cut right after it (04, 08, 09, 13, 14).

Does SIG ask brain teasers in the quant research intern interview?

Mostly probability rather than pure brain teasers: logic puzzles came up in 3 of 20 current PhD reports (07, 15, 16), and classic questions like the coin pattern race and the bus question recur (Patterns, point 2).

Is there a data exercise in the SIG intern process?

For candidates who get past the first calls, yes: a data round or exercise appears in 01, 07 and 18, and older reports describe a week-long exercise (51, 57).

Keep practicing

Sources: Anywhere useful interview info can be found.

0 total reactions. Your reactions: none.

Comments

0

No comments yet — be the first to share what you know.