Blog6 min read
How to prepare for a quant interview
What each round tests, and a plan you can start this week.
The short answer
Quant interviews test a few things over and over: probability and expected value, quick and careful mental arithmetic, how you reason when you're unsure and, depending on the role, statistics, coding or how markets work. Most hiring loops start with an online test, then phone or video screens, then a final round. Prepare the shared core first, then the part your role leans on, and practise out loud: interviewers score how you think as much as the answer.
Three roles, one core
What traders, researchers and developers are each tested on.
Quant firms hire for three main jobs. Traders set prices and decide what to buy and sell. Researchers look for patterns in data that predict prices, and turn them into strategies. Developers build the systems that trade: fast, correct and hard to break. Each job is tested differently, so know which one you're preparing for.
| Role | Leans on | Also tested |
|---|---|---|
| Quant trader | Mental maths, expected value, making markets, betting games | Probability, quick estimates |
| Quant researcher | Probability, statistics, a research or modelling case | Coding with data, how you weigh evidence |
| Quant developer | Coding, data structures, C++ and performance | Some probability, designing systems such as an order book |
Not sure which suits you? Quant trader vs quant researcher and quant researcher vs quant developer set the jobs side by side.
The rounds
From the online test to the final day, and what each stage looks for.
- Online test. Timed and automatic. Expect fast arithmetic, number sequences, probability questions or a coding task, depending on the role. Some tests take marks off for wrong answers, so read the instructions before you start.
- Phone or video screens. One or two interviews of 30 to 60 minutes: probability puzzles, expected value, mental maths and questions about your background. You talk while you work.
- Role rounds. Traders get market-making games and betting questions. Researchers get statistics and a case: here is some data, what would you test? Developers get coding, design and performance questions.
- Final round. Several interviews in a row, mixing all of the above with questions about why you want the job and how you work with others.
insideout's role pages map these rounds in detail, with practice for each: quant trader, quant researcher and quant developer.
Probability and expected value
The topic every quant interview shares.
Probability is the one topic every quant interview shares. A small toolkit covers most questions: count outcomes carefully, condition on what you know, use expected value, and break a hard problem into a first step and what follows.
Example 3.1
A die you may roll again
You roll a fair die and are paid its face value in dollars. After seeing it, you may roll once more and take the second result instead. What is the game worth?
- Value of a fresh roll(1 + 2 + 3 + 4 + 5 + 6) ÷ 6 = 3.5
If you roll again, you take whatever comes up: $3.50 on average.
- Decide when to keep the first rollkeep 4, 5 or 6
Keep it when it beats 3.5. Roll again on 1, 2 or 3.
- Average when you keep(4 + 5 + 6) ÷ 3 = 5
Half the time you keep, and then you average $5.
- Combine the two cases1/2 × 5 + 1/2 × 3.5 = 4.25
Answer$4.25. The chance to roll again is worth 75 cents more than a single roll.
Example 3.2
A positive test
2% of people have a condition. A test catches it 90% of the time, and wrongly flags 5% of people who don't have it. Someone tests positive. How likely is it that they have the condition?
- Picture 1,000 people20 have it, 980 don't
Counts are easier to reason with than percentages.
- Count true positives90% × 20 = 18
- Count false positives5% × 980 = 49
A small error rate on a big group makes many false alarms.
- Share of positives who have it18 ÷ (18 + 49) = 18 ÷ 67 ≈ 0.27
AnswerAbout 27%. Most positives are false alarms, because the condition is rare.
Probability questions in quant interviews works through ten more, and Probability teaches the methods from the start.
Mental maths
Fast, careful arithmetic, and the habits that keep it right.
Trading roles, and many online tests, expect quick arithmetic without a calculator: multiplication, fractions, percentages and rough estimates. It is a trained skill. Ten to fifteen minutes a day for a few weeks makes a large difference.
Example 4.1
Two shortcuts that save seconds
Work out 48 × 52 and 7% of 340 in your head.
- Write 48 × 52 as a difference of squares(50 − 2) × (50 + 2) = 2,500 − 4 = 2,496
It works for any two numbers spaced evenly around a round number.
- Split the percentage1% of 340 = 3.4, so 7% = 7 × 3.4 = 23.8
Answer2,496 and 23.8.
Check the size of every answer before you say it: 48 × 52 must land a little under 50 × 50. The mental maths test, free with an account, runs timed sets with fresh numbers, and mental maths for trading interviews covers the methods.
Making a market
How to quote a price on something uncertain.
Trading interviews often include market-making games. The interviewer names something uncertain and asks for a market: a price you would buy at, the bid, and a price you would sell at, the ask. Then they trade with you and change the facts, to see how you update.
Example 5.1
A market on two dice
Make a market on the sum of two fair dice, to be rolled after you trade.
- Find the fair value3.5 + 3.5 = 7
The expected sum is the sum of the expected values.
- Size the uncertainty√(2 × 35 ÷ 12) ≈ 2.4
One die has variance 35 ÷ 12. Two out of three rolls land between 5 and 9.
- Quote around the fair value6.5 bid, 7.5 ask
A tight spread shows you know the value. Widen it when you're less sure.
- Update when you learn somethingone die shows 6: 6 + 3.5 = 9.5
Move the whole market at once, then quote around the new value.
Answer6.5 at 7.5 to start, centred on 7.
Interviewers care less about your exact numbers than about whether the quote is centred, the spread is sensible, and you update quickly when the facts change. Market Microstructure explains how real order books and market makers work.
An eight-week plan
About an hour a day, with the shared core first.
| Weeks | Focus | Each day |
|---|---|---|
| 1 and 2 | Counting, probability rules, conditional probability | One lesson, five problems, 10 minutes of mental maths |
| 3 and 4 | Expected value, Bayes' rule, common distributions, first-step analysis | Five problems out loud, 10 minutes of mental maths |
| 5 and 6 | Your role's focus: markets and betting for traders, statistics and a case for researchers, coding and systems for developers | One role topic, one timed set |
| 7 | Mock rounds under time, out loud | One full round, then go through every mistake |
| 8 | Weak spots, your story, why this job | Short sets only, and rest the day before |
Keep a list of every problem you got wrong and redo it three days later. The mistakes you repeat tell you what to study next.
How you talk matters
What interviewers score besides the answer.
- Say your plan before you calculate. "I'll condition on the first roll" lets the interviewer follow you, and help if you drift.
- Check sizes. A probability above 1, or an average wait shorter than one try, means a slip. Catch it before they do.
- Say when you're unsure. Give your best estimate and how you would tighten it. A confident guess that's wrong does more harm.
- Use hints. A hint is part of the interview. Take it and keep going.
- Know why you want the job. Most hiring loops ask. Have a short, true answer.
Interview on insideout lets you pick your role and your date, then practise each round in order.
Learn it properly
- Counting outcomesProbability · Lesson 01 Free
- Probability is countingProbability · Lesson 02 Free
- Conditional probabilityProbability · Lesson 03 Free
- What an exchange isMarket Microstructure · Lesson 01 Free
- Samples and estimatesStatistics and Regression · Lesson 01 Free
- A puzzle you've never seenPuzzles and Brainteasers · Lesson 01 Free
Questions
How long does it take to prepare for a quant interview?
With a maths, physics or computer science background, six to eight weeks at about an hour a day is a common plan. Start sooner if probability is new to you, and keep mental maths going every day.
Do I need a finance background?
No. Most quant interviews test maths, reasoning and coding, and teach the finance on the job. Knowing how an order book, a bid and an ask work helps in trading rounds.
Which programming language should I use?
Python is the usual choice for research and data rounds. Developer roles often expect C++ as well. Use the language you know best unless the role says otherwise.
What if I don't know the answer?
Say what you would try, solve a simpler version first, and give an estimate. Interviewers want to see how you reason when the path isn't obvious.