SIG Interview 2026: OA, HackerRank, Superday & Poker

SIG Interview Process at a Glance

Candidates report the whole process taking two to eight weeks, most often two to four.

---

What to Expect from a SIG Interview

Susquehanna International Group, almost universally called SIG, is one of the most distinctive interview processes in trading. The firm built its reputation on probabilistic thinking - their internal training programme famously uses poker as a teaching tool for game theory and decision-making under uncertainty - and the interview reflects that culture exactly. You will be asked questions where the correct answer matters less than how you reason about expected value, edge, and information.

This guide walks through the full SIG process for trader, quant and developer roles, the question types that come up most often, and a six-week preparation plan you can run in parallel with the Quantt coding tests. For the broader landscape of comparable employers, see our quant interview questions guide which covers the wider field.

---

SIG at a Glance

• Founded: 1987 by Jeff Yass and his Wharton classmates
• Headquarters: Bala Cynwyd, Pennsylvania (with major offices in London, Dublin, Hong Kong and Sydney)
• Size: 2,500+ employees globally
• What they trade: Virtually every listed equity options product in North America, plus equity, fixed income, commodity and ETF market making worldwide
• Roles they hire: Quantitative Trader, Quantitative Researcher, Software Developer, Quantitative Systematic Trader
• Application route: SIG careers page or via the Susquehanna firm page on Quantt

For broader context, our prop trading firms guide compares SIG to Optiver, IMC, Jane Street and others in the options market-making space.

---

The Interview Process - Stage by Stage

SIG publishes very little about its process. The one official flow we found is in its Equity Research internship ad: "Apply > Quantitative & Finance Assessment > Recruiter Interview > Interview > Final Interview > Offer". The trader track looks similar in candidate reports. Everything below that is not marked official comes from candidates and prep sites.

Stage 1: The Online Assessment

What SIG says. Its job ads state that "applications for this job require an online assessment", administered by "Mercer | Mettl", and that by applying you consent to your email address being passed to Mettl so it can contact you directly. SIG publishes no test length, question count or pass mark.

That wording suggests the invite goes out automatically when you apply. Some prep sites say it follows a CV screen instead, and one candidate who scored 17 out of 17 still did not get an interview, which hints that the CV is read after the test. Nobody has confirmed the order.

The Problem Solving Assessment (PSA). For the 2026/27 cycle, trader candidates report a 60-minute test of 17 questions. At least six sources give that format, including several Reddit posts from June to August 2026, one of which found it "harder than the sample questions". The topics reported are:

• Probability, including continuous probability
• Expected value and combinatorics
• Bayes' rule and Markov chains
• Single-variable calculus
• Deductive logic puzzles

Prep sites describe the conditions: a calculator, pen and paper allowed; answers entered as integers or fractions, with decimals possibly marked wrong; the option to move back and flag questions; a test that can end if you switch tabs; and seven days to start it. No primary source confirms those details, so read the instructions on your own invite.

The older Quantitative Evaluation. Until recently, many trader candidates got a shorter test: about 20 minutes, with question counts reported between 12 and 16, covering probability, arithmetic, rates and logic. One August 2026 post says SIG "changed this year from the 20 min quant evaluation to the 60 min psa for qt roles", and JobTestPrep says the shorter test is no longer used. Some candidates still describe the 60-minute test as the research-role version and the 20-minute one as the trader version, so which you get may depend on the role.

Which platform. SIG's ads still name Mettl. Two August 2026 Reddit posts say their PSA ran on HackerRank, while a June 2026 thread says Mettl. Expect either.

Two questions in the reported PSA style (our examples, not leaked questions):

• A point is chosen uniformly in the unit square. What is the probability that x + y < 1.5? The excluded region is the corner triangle above the line, with legs of 0.5, so area 0.125. The answer is 7/8. Give it as a fraction.
• What is the expected number of fair coin flips to get two heads in a row? Let E be the expected flips from scratch. After a tail you restart; after heads then tails you restart; after two heads you stop. E = 0.5(1 + E) + 0.25(2 + E) + 0.25(2), which gives E = 6.

Developer Coding Tests

For software roles, most candidate reports name CodeSignal. A Dublin intern invite in December 2024 came through CodeSignal, with replies mentioning a 2D-array question and a two-pointer problem. Glassdoor reviews describe "60 min CodeSignal Python test… 3 questions" and, for a new-grad role, four questions in two hours. One Glassdoor review describes two easy problems in 90 minutes. Expect problems on arrays, hash maps and simulation graded on hidden test cases, so edge cases and complexity matter as much as a working answer.

The SIG Data Exercise (Quant Research)

Some quantitative research candidates receive a take-home data exercise instead of, or after, the timed test. Candidates describe a dataset of market-style or synthetic features, a prediction target, and a request to explore the data, build a model and write up the findings. What gets marked is judgment: sensible cleaning, honest out-of-sample validation, a clear explanation of what you would try next, and no overfitting. A simple, well-validated model with a clear write-up beats a complex one you cannot defend.

How to Prepare for the OA

• Work through quant probability interview questions until conditional probability, expected value and Markov chain set-ups are automatic - that is the core of the PSA
• Practise giving exact answers as fractions, and revise single-variable calculus for continuous probability
• Keep up arithmetic speed on the mental maths trainer for the live rounds, where you will not have a calculator
• Read our quant online assessments guide for how SIG's test compares with Optiver, IMC and Jane Street

Stage 2: The Recruiter Call

A short phone call covering your CV and a handful of probability and expected value questions. A 2022 write-up of the Dublin trader programme describes five probability questions on this call. One June 2026 candidate was invited to it about two days after finishing the online test.

Stage 3: Technical Interviews

For traders, candidates describe one or two video interviews of about 60 minutes with a senior trader, built around probability, expected value and short games. For software candidates, Glassdoor reports describe a behavioural round with a recruiter, then live coding of about 30 minutes.

Stage 4: The Superday

For traders, this is where SIG's culture shows. Candidates describe expected value and market-making games played in person with cards, dice and coins, a group auction or bidding game, and an HR interview of about 30 minutes. For software candidates, the final stage is reported as an onsite coding round plus low-level design.

The games are the part to practise. You are making markets and sizing bets in real time, without a calculator, while someone watches how you update. Our market making interview questions work through the formats.

Where Poker Really Fits

Poker is famous at SIG, but as training rather than as an interview round. Tradermath's SIG guide says poker belongs to post-hire culture and training and that, as a candidate, "you will not be dealt in". We found no recent candidate report of a poker round during interviews. What does carry over is the reasoning poker teaches: pot odds, expected value, Bayesian updating and the difference between variance and edge. That is what the Superday games test.

SIG is competitive on compensation. Graduate quantitative traders are reported to receive a first-year package between $200,000 and $300,000 in the US, with comparable packages in Dublin and London. These are estimates, not SIG figures.

---

Did you know?

Tap a tile for a quick interview snippet — then keep going.

• Why SIG loves probability + poker

  Susquehanna’s culture blends options market making with game-theory and decision-making under uncertainty. EV, Bayesian updating, and “what’s your edge?” language show up a lot.

• Mental maths still matters

  Alongside puzzles, expect quick arithmetic and rough odds — not always Optiver-extreme timing, but fluency under conversation pressure.

• SIG vs Jane Street vibe

  Both are elite prop shops. SIG often feels more market-maker / poker-flavoured; Jane Street more academic / engineering. Prep the shared core: probability, EV, clear talk-alouds.

How Interviews Differ by Role

Quantitative Trader

This is the role most associated with SIG. The interview leans heavily on probability, expected value and decision-making under uncertainty. Mental maths matters enormously - candidates report being asked to multiply two-digit numbers or compute percentages within five seconds throughout the day.

Typical question split:

• 35% probability and expected-value puzzles
• 25% mental maths and estimation
• 25% market intuition and pricing
• 15% behavioural and culture fit

Quantitative Researcher

More statistical and modelling-heavy than the trader track. Expect questions on hypothesis testing, regression, time series, and the design of statistical signals. Coding (Python) appears in most interviews, usually focused on data manipulation rather than algorithmic puzzles.

Typical split:

• 40% statistics and probability theory
• 30% Python and data manipulation
• 20% modelling and signal design
• 10% behavioural

Software Developer

Closer to a FAANG-style technical interview, with a finance flavour. SIG's technology stack runs on C++ and Python, so fluency in at least one is non-negotiable. Expect algorithm and data structure questions, plus systems design questions specific to trading infrastructure - order books, low-latency message passing, market data normalisation.

Typical split:

• 50% coding (algorithms and data structures)
• 25% systems design
• 15% language-specific knowledge (C++ memory model, Python performance)
• 10% behavioural

Quantitative Systematic Trader

A hybrid track that sits between the trader and research roles. Interview content covers all three: probability puzzles for the trader instinct, statistical modelling for the research dimension, and Python coding for the implementation side.

---

Real Question Types

The questions below are representative - we have collected them from candidates and from SIG's own published practice materials. Your exact questions will differ but the shape will not.

Probability and Expected Value

Question 1: Two coins You have two coins. One is fair. The other is biased and lands heads with probability 0.75. You pick a coin at random and flip it three times, getting three heads in a row. What is the probability you picked the fair coin?

Approach: Bayes' theorem. P(fair | HHH) = P(HHH | fair) * P(fair) / P(HHH). The numerator is 0.5 * 0.5 = 0.0625. The denominator is 0.5 * 0.125 + 0.5 * 0.421875 = 0.273. The answer is roughly 22.9%. The interviewer wants to see you set up Bayes correctly, not just guess.

Question 2: Expected dice value with reroll I roll a fair six-sided die. You can either accept the roll or reroll once. What is your expected value if you play optimally?

Approach: You should reroll if and only if your first roll is below the expected value of a fresh roll, which is 3.5. So you reroll on 1, 2 or 3 (probability 0.5) and keep on 4, 5 or 6 (probability 0.5). E = 0.5 * 3.5 + 0.5 * E[X | X >= 4] = 0.5 * 3.5 + 0.5 * 5 = 4.25.

Brain Teasers and Estimation

Question 3: Estimation How many AAA batteries are sold in the United Kingdom each year?

Approach: Top down. UK population about 67 million. Estimate households (~28 million), average AAA-using devices per household (~3 to 5), battery replacement frequency (~once per year per device). That gives roughly 28 million * 4 * 1 = 112 million household batteries. Add commercial and institutional use (call it another 30 to 50%), and you arrive at 150 to 170 million per year. The number is not the point; the structure is.

Question 4: Trading game I will pay you the sum of three rolls of a fair six-sided die. How much would you pay to play, and at what price would you flip and become the seller?

Approach: Expected value is 3 * 3.5 = 10.5. You would buy at any price below 10.5 and sell at any price above it. The follow-up is usually about variance and risk aversion - the standard deviation of the sum is sqrt(3 * 35/12) ≈ 2.96, so a risk-averse trader would shade their bid down and their ask up.

Coding Questions

Question 5: Two-pointer maximum subarray product Given an array of integers (positive, negative and zero), return the maximum product of any contiguous subarray. Aim for O(n) time and O(1) space.

def max_product_subarray(nums: list[int]) -> int:
    if not nums:
        return 0
    max_prod = min_prod = result = nums[0]
    for x in nums[1:]:
        candidates = (x, x * max_prod, x * min_prod)
        max_prod, min_prod = max(candidates), min(candidates)
        result = max(result, max_prod)
    return result

The trick is tracking both the running maximum and minimum, because multiplying by a negative flips signs. Interviewers will probe edge cases (single element, all zeros, single negative) and ask you to argue complexity.

Question 6: Order-book matching Design a class that supports add_buy(price, qty), add_sell(price, qty) and match() returning a list of trades. Limit orders only, price-time priority.

Approach: Use two sorted structures (heap or balanced tree) - a max-heap on the bid side and a min-heap on the ask side. On every add, check whether the best bid crosses the best ask and pop fills until the spread reopens. Time-priority is handled by tagging each order with an incrementing sequence number.

Market Knowledge and Pricing

Question 7: Put-call parity A non-dividend-paying stock trades at $100. A 90-day at-the-money call is priced at $5. The risk-free rate is 4% annualised. What does put-call parity say the put is worth?

Approach: C - P = S - K * exp(-rT). With S = K = 100, r = 0.04, T = 90/365, the right-hand side is 100 - 100 * exp(-0.04 * 90/365) = 100 - 99.014 ≈ 0.986. So P = C - 0.986 ≈ $4.01.

Behavioural

Question 8: Why SIG? Tell me why SIG over Jane Street, Optiver or IMC.

Approach: Generic answers fail. Specific things to point at: SIG's training programme is unusually structured and explicit (they teach poker, decision theory and game theory formally), the firm has a reputation for keeping traders for long careers rather than burning them out, and their options business is genuinely the broadest in North America. Tie one of those to something specific you have done or want to do.

---

Quick quiz

SIG-style probability check

Expected value and conditional thinking.

1 / 3

A game pays $10 with probability 0.4, else $0. What’s the EV?

• $2
• $4
• $6
• $10

How to Prepare - A Six-Week Plan

This is the schedule we suggest to Quantt members preparing specifically for SIG.

Week 1 - Mental maths

Daily 30-minute drills using Zetamac or the free Quantt mental maths trainer (two-digit by two-digit multiplication, three-digit by one-digit division, and percentages). The current online test reportedly allows a calculator, so this week is for the live rounds and games, where you work without one.

Week 2 - Probability foundations

Work through chapters 1 to 7 of A Practical Guide to Quantitative Finance Interviews by Xinfeng Zhou (the "green book" - see our Green Book study guide), then test yourself on our quant probability interview questions. Pair this with our probability for quant finance and statistics for quantitative trading guides for the underlying theory. By end of week 2 you should be comfortable with conditional probability, expected value, Bayes' theorem and the standard distributions.

Week 3 - Brain teasers and game theory

Work through Heard on the Street by Timothy Crack (our review covers how to use it), our quant brain teasers, and the puzzle archive at Quant Net. Read or skim The Mathematics of Poker (Chen and Ankenman) to understand expected value in adversarial settings. You do not need to memorise poker theory - you do need the intuition.

Week 4 - Coding

If you are interviewing for a trader role, focus on Python data manipulation and one or two algorithmic problems per day. For developer roles, switch to LeetCode (medium and hard) plus a daily systems design exercise. The Quantt coding tests are calibrated to the type of problems SIG asks - they are a good way to time-box your prep.

Week 5 - SIG-specific patterns

Run mock Superdays. Three back-to-back 45-minute sessions with a study partner, alternating between probability, market-making games with cards or dice, and coding. Time yourself ruthlessly. A long day of games is tiring, and the only way to prepare for that is to practise tired.

Week 6 - Polish and behavioural

Refine your answers to "why SIG", "tell me about a time you took a calculated risk", and "tell me about a mistake you made". Read SIG's recent press and any interviews with senior leadership (Jeff Yass occasionally appears in podcasts). Salary research using Glassdoor and the Quantt quantitative analyst salary guide.

Recommended Reading

---

What SIG Looks For Beyond Technical Skill

The technical bar is high but the differentiating signal is rarely pure intelligence. Three traits separate offers from rejections at the Superday stage.

Calibration. SIG cares more about whether you know what you do not know than whether you know more. A candidate who answers "I am 70% confident the answer is roughly 4.2, and here is why" usually beats one who answers "the answer is 4.2" with no uncertainty quantification. This is a direct legacy of the firm's poker culture, where overconfidence costs money in measurable ways.

Curiosity under questioning. Interviewers will push back on your answers, sometimes when you are right. They are testing whether you treat new information as a signal to update or an attack to defend against. The candidates who do well treat the pushback as data.

Genuine interest in markets. SIG hires people who would think about expected value problems for fun. If you cannot name a market structure question that genuinely interests you, the firm is unlikely to be the right fit, and the interviewers will pick up on that quickly.

For more on the broader skills firms across the quant industry are looking for, see our quantitative analyst career guide.

---

Compensation & recruiting notes

Pay ranges in this guide are illustrative estimates from publicly discussed bands and anecdotal reports - not official figures from the employer. Packages vary widely by desk, office, performance, and year. Hiring processes change; nothing here guarantees an interview, assessment format, or offer.

---

Frequently Asked Questions

How hard is it to get a SIG interview?

Very competitive, though SIG does not publish hiring numbers or pass rates. The strongest route in is through campus recruiting at target universities (Cambridge, Oxford, LSE, Imperial, Warwick in the UK; MIT, Princeton, Penn, Carnegie Mellon, NYU in the US) or via referral from a current employee. Online assessment scores are taken seriously regardless of school - many candidates from non-target universities make it through on OA performance alone.

Does SIG hire from non-target universities?

Yes. SIG's online assessment is genuinely meritocratic and many of their best traders did not come from Ivy League or Oxbridge backgrounds. The realistic path for a non-target candidate is to absolutely crush the OA and have one or two demonstrable signals - a strong showing on Codeforces or ICPC, a competitive poker or trading record, or a published research project.

What programming languages should I know for a SIG interview?

For trader roles, Python is the only language tested and even then only at a basic data-manipulation level. For quant research roles, Python plus enough R or MATLAB exposure to read code in those languages. For software developer roles, C++ is heavily emphasised - expect questions on memory management, move semantics and performance characteristics. Familiarity with kdb+/q is a significant plus for some teams.

How does SIG's compensation compare to other quant firms?

SIG sits in the upper tier but typically below the very top (Citadel, Jane Street and Hudson River). Graduate quantitative traders in the US receive a first-year total package between $200,000 and $300,000, including signing bonus and first-year guarantee. In the UK and Ireland, the equivalent is roughly £130,000 to £180,000. Senior traders and partners earn substantially more, with the firm's profit-sharing structure giving high upside for top performers. For a comprehensive breakdown across firms see our quantitative analyst salary guide.

What is on the SIG online assessment?

For trader and research roles in the 2026/27 cycle, candidates report the Problem Solving Assessment: 17 questions in 60 minutes on probability, expected value, combinatorics, Bayes, Markov chains, calculus and logic, with a calculator reportedly allowed. An older 20-minute Quantitative Evaluation is now reported far less often. For developer roles, a coding test, most often reported on CodeSignal. SIG's ads name Mercer Mettl as the assessment provider.

What is the SIG data exercise?

A take-home task given to some quantitative research candidates: explore a dataset, build a predictive model and write up your findings. SIG is assessing validation discipline and communication more than model complexity.

What happens at the SIG Superday?

For traders, candidates describe expected value and market-making games played with cards, dice and coins, a group auction or bidding game, and an HR interview of about 30 minutes. For software candidates, an onsite coding round plus low-level design.

How many interview rounds does SIG have?

Candidates describe four stages for traders: the online assessment, a recruiter call, one or two technical video interviews, and a Superday. Reports put the whole process at two to eight weeks, most often two to four.

Do I really need to know how to play poker for SIG?

No. Poker is part of SIG's training and culture after you join, and one detailed prep guide says candidates "will not be dealt in" during interviews. You do need the reasoning poker teaches (pot odds, expected value, Bayesian updating and sizing bets to your edge), because the Superday games test exactly that.

Can I reapply if I'm rejected?

SIG does not publish a reapplication rule, so check with the recruiter or the careers page for your cycle. If you were rejected at the online test, the reported PSA format (17 probability, logic and calculus questions in 60 minutes) rewards depth on conditional probability and Markov chains more than raw arithmetic speed.

What is the difference between SIG and Jane Street?

Both are options-heavy proprietary trading firms with strong probability cultures, but they diverge sharply on technology and culture. Jane Street is OCaml-everything with a famously academic and engineering-driven culture. SIG is more market-maker-traditional, uses C++ and Python, and leans much more heavily on poker and game theory in training. Compensation is broadly similar at entry level. We have a dedicated Jane Street interview guide covering Jane Street's process in detail.