Jane Street Internship: Trading, Quant Research & Strategy Interviews

Jane Street's summer internships run for 10 to 12 weeks between May and September in New York, London, and Hong Kong across Quantitative Trading, Quantitative Research, Strategy and Product, Software Engineering, and Machine Learning. According to Jane Street's 2026 New York job postings, the annualized base salary is $300,000 for the Quantitative Trader Internship and $250,000 for the Strategy and Product Internship. You do not need a finance background, a minimum GPA, a cover letter, or prior experience with OCaml to apply, and a single application is automatically evaluated for all open roles globally.

Updated

Internship tracks and New York base pay

Jane Street divides its core analytical and market-facing summer internships into several tracks. A real human reviews every application, and recruiters automatically consider you for all open roles globally to find the best match.

If you want to warm up before applying, you can practice a case free, no sign-up to test your structured thinking and mental math out loud.

Jane Street summer internship tracks and published requirements

Jane Street summer internship tracks and published requirements
TrackCore summer workInterview focusPrior coding or finance needed?
Quantitative TradingPaired with two trader mentors on desk projects, classes, mock trading, and one summer electiveTrader-led phone screens and in-person final rounds on probability, problem solving, coding, and data analysisNo finance background required. General coding experience is a plus, in any language
Quantitative ResearchBuilding mathematical models in code across sub-microsecond to long-term horizonsA mix of Quantitative Trading and Software Engineering interviewsNo finance or OCaml required. Code in the language you know best
Strategy and ProductAnalytical and cross-functional business, operational, and product projects with full-time mentorsOne-hour online assessment, one or two video scenario interviews, and an in-person final dayNo prior financial knowledge, specific major, or coding experience required
Software EngineeringBuilding in-house trading, network, and platform systemsZoom technical coding round followed by in-person collaborative coding roundsStrong coding in your best language required. No OCaml or finance needed

Inside the three core tracks: Trading, Research, and SP

In the Quantitative Trading internship, you are paired with two different experienced traders on projects tied to their daily desk work. Past intern projects have included analyzing datasets, training predictive models, simulating new trading strategies, and writing production tools. You have access to petabytes of data, a computing cluster with hundreds of thousands of CPU cores, and a GPU cluster with tens of thousands of high-end GPUs. Alongside classes and team mock trading sessions, every trading intern chooses one summer elective.

  • Machine Learning, Modeling, and Data Science (Modeling Elective): You train statistical and machine learning models on real and simulated market data, learning how noisy financial data differs from textbook datasets.
  • Algorithmic Trading and Market Microstructure (Algo Elective): You analyze market data to build a tradable fair value and implement an automated trading strategy in Python that connects to simulated markets.
  • Trading Strategy and Scenarios (Strategy Elective): You work in teams across rotating mock trading scenarios inspired by real historical events on a trading desk, refining strategies between sessions and reviewing what actually happened with the traders who lived through it.

Quantitative Research and Strategy and Product work

In Quantitative Research, you act as part trader and part engineer. Researchers collaborate with trading and machine learning teams to develop novel mathematical models and implement them in code for automated trading strategies. Their problems range from sub-microsecond trading to finding long-term market inefficiencies across trillions of historic events.

In Strategy and Product (SP), interns work as analytical generalists solving cross-functional problems across trading, technology, finance, compliance, and operations. Full-time SPs start with a year-long rotation across the firm, and interns mirror that work by breaking down business constraints, modeling capacity with external providers, and designing workflows for new products or regulatory shifts like the move from T+2 to T+1 settlement.

The interview process step by step

After you submit your resume (no cover letter is needed), a recruiter reviews it and typically responds within a few days to a week. If you already hold competing offers or exploding deadlines, tell your recruiter so they can expedite your schedule. Interviews run Monday through Friday during local business hours, never on weekends. When you reach the final in-person stage, Jane Street books and pays for all of your flights and hotel accommodations. You will hear back within a week of your interviews whether or not you are moving forward.

Quantitative Trading interviews start with phone interviews (or Zoom if you notify them ahead of time about a poor phone connection) and finish with an in-person final round. Every interview is conducted by quantitative traders. They do not test finance, economics, or higher-level college math. Instead, the rounds work like a conversation covering problem solving, probability and statistics, coding in any language you choose, data analysis, and your general interests.

Quantitative Research interviews combine the Quantitative Trading and Software Engineering processes. You will face probability and market-making questions alongside live coding challenges in your strongest programming language. Jane Street asks that you write real code rather than pseudocode, and strongly advises against trying OCaml for the first time in an interview.

Strategy and Product interviews follow a distinct three-stage path designed to require no outside preparation. Second, you do 1 or 2 video interviews with a full-time SP, working through longer-form hypothetical business scenarios and logic puzzles in real time. Third, you attend a full day of in-office interviews that explore deeper qualitative and technical problems representative of daily SP projects.

Probability, expected value, and confidence intervals

Jane Street publishes an official Probability and Markets guide covering the exact building blocks tested in trading and research interviews. Interviewers look for intuitive reasoning under uncertainty rather than memorized theorems.

When outcomes are finite and equally likely, you find probabilities by counting the outcomes that fulfill your criteria and dividing by the total number of outcomes. For two six-sided dice (2d6), treating the dice as distinct (such as one red and one blue) gives 36 equally likely outcomes. A sum of 7 can happen in 6 ways ([1,6], [2,5], [3,4], [4,3], [5,2], [6,1]), giving a probability of 6/36, or 1/6.

When you attach a numerical payout to outcomes, you create a random variable X. Its average outcome is the expected value, written as E[X], calculated by summing each outcome multiplied by its probability. For a single six-sided die (d6), E[X] is 3.5.

  • Linearity of expectation: For any two random variables X and Y, E[X + Y] = E[X] + E[Y]. This holds even when X and Y are not independent, which simplifies multi-part interview problems.
  • Product of independent variables: If X and Y are independent, E[X * Y] = E[X] * E[Y]. If a number c is constant, E[c * Y] = c * E[Y].
  • Indicator functions for betting: Assign 1 if event A happens and 0 if it does not (written as 1_A, where E(1_A) = P(A)). If a bet pays x when A happens and y (typically negative) when it does not, your expected payout is y * P(not A) + x * P(A), which simplifies to y + (x minus y) * P(A).
  • Conditional probability and updating priors: As new information arrives, you update your prior belief using P(A given B) = P(A and B) / P(B). If someone rolls a d6 and tells you the result is at least 4, the remaining outcomes (4, 5, 6) each have a 1/3 probability, shifting the expected value from 3.5 up to 5.
  • Recursion for memoryless games: If a process has the same distribution regardless of earlier trials, set up a recursive equation. To find the expected number of coin flips to see the first heads, note that you spend 1 flip, succeed half the time (0 more flips), and restart half the time: E(flips) = 1 + (1/2)*(0) + (1/2)*E(flips), which solves to E(flips) = 2.
  • Confidence intervals and calibration: Interviewers ask for confidence intervals, such as a 90% or 95% range around an unknown quantity (for example, the number of taxis in New York City). People tend to be overconfident and give intervals that are too narrow, so practice setting wide enough bounds that the true number falls inside your range 95% of the time.

Market making games and adverse selection

In trading interviews, you connect expected value to live market-making games. To place a trade, you specify four things: the contract, the direction (buy or sell), the price, and the quantity (size). A price to buy is a bid, and a price to sell is an offer.

When you make a market around a contract's expected value, you buy below E[X] and sell above E[X]. For a contract that pays $6 per pip on a d6 roll (expected value $3.5), bidding $3.5 breaks even on average, while bidding $1 will likely get no trades. Balancing profit per trade against the likelihood of getting filled, a reasonable market is "3 at 4, 10 up." However, if the contract pays $1,000,000 per pip, rolling a 1 on a $3,500,000 purchase loses $2,500,000. When a single loss can wipe out a large portion of your capital, you must widen your spread or reduce your size to manage risk.

You also have to account for adverse selection. You are trading against a human who is selecting into the trade against you. If you make a market and your counterparty immediately hits your bid or lifts your offer, ask yourself why. If they are trying to maximize expected value on a complex estimate (like trading weather with a meteorologist), your order getting filled is a signal that your price was too generous. You must update your belief conditional on your order being filled: if someone sells to your bid, the true value is likely lower than you initially calculated.

  • "I'm 2 bid for 10": You are willing to buy 10 units at $2 each.
  • "I have 10 at 4": You are willing to sell 10 units at $4 each.
  • "I'm 2 at 4, 10 up": You are willing to either buy 10 units at $2 or sell 10 units at $4. Your order stays active until someone trades against it or you cancel it by saying "I'm out."
  • "Sold": What a counterparty says to hit (sell to) your entire bid quantity.
  • "Take 'em": What a counterparty says to lift (buy) your entire offer quantity. For a partial fill of 5 units, they say "I buy 5" or "I sell 5," leaving the remaining 5 active.

Worked example: pricing a two-roll die contract

Jane Street's Probability and Markets guide asks how to make a market on a d6 roll where the holder has the choice to reroll once if they do not like the first roll, keeping the second roll no matter what. Here is our team's worked example showing how to solve and quote this contract step by step.

Step 1: Clarify who holds the option. In this contract, the buyer (holder) decides whether to reroll after seeing the first die. Assuming a rational holder who wants to maximize expected payout, they will only keep the first roll if it is strictly better than the expected value of rolling again.

Step 2: Find the expected value of the second roll. Because the second roll is just a single standard d6, its expected value is 3.5.

Step 3: Determine the optimal strategy on the first roll. If the first roll is 4, 5, or 6, the payout is above 3.5, so the holder keeps it. If the first roll is 1, 2, or 3, the payout is below 3.5, so the holder throws it back and takes the second roll (which has an expected value of 3.5).

Step 4: Calculate the overall expected value. On the first roll, there is a 1/2 chance of rolling 1, 2, or 3 (yielding an expected 3.5 from the reroll), a 1/6 chance of rolling 4, a 1/6 chance of rolling 5, and a 1/6 chance of rolling 6. Summing these gives E[X] = (3/6)*(3.5) + (1/6)*(4) + (1/6)*(5) + (1/6)*(6) = (10.5 + 15) / 6 = 25.5 / 6 = 4.25.

Step 5: Quote a two-way market. Since the fair value is $4.25 and the outcomes are bounded between $1 and $6 with known probabilities, adverse selection is low (unless you miscalculated). To earn an edge while encouraging trades, you can quote a market straddling 4.25, such as "I'm 4 at 4.50, 10 up."

How to prepare: a 4-step checklist before you apply

Jane Street notes that common internet rumors about its interviews are often false. You will not be judged harshly if you cannot multiply double-digit numbers in 5 seconds, and you do not fail automatically if you make a mistake. Interviewers want to see how you think out loud, check order-of-magnitude edge cases (like testing 0, 1, or 2 players before solving for 100 players), take hints, and correct errors with humility. Use this checklist to build those habits before submitting your application.

  • 1. Master the four core probability tools: Practice counting distinct outcomes, applying linearity of expectation E[X + Y] = E[X] + E[Y], writing recursive equations for memoryless games, and updating priors with P(A given B) = P(A and B) / P(B).
  • 2. Run calibration drills on everyday numbers: Pick 10 verifiable facts (such as city populations or geographic distances) and write down 90% and 95% confidence intervals for each. Check how many true answers land inside your bounds to fix overconfidence.
  • 3. Practice quoting markets out loud: Take standard dice and card problems, compute the fair value, and practice saying "I'm X at Y, 10 up" out loud. Then simulate what happens to your quote if your interviewer says "Take 'em" three times in a row.
  • 4. Rehearse collaborative problem solving: Practice talking through unfamiliar logic puzzles and business scenarios out loud, asking clarifying questions early, and stating when you are unsure rather than guessing.

Practice quant cases and problem solving out loud

Reading probability formulas on a page feels very different from explaining your reasoning while an interviewer changes the rules or trades against your quote. You can use Soreno to practice brain teasers and quant cases, structured cases, quantitative estimation, and follow-up pressure out loud with an AI interviewer, or see a real sample report to check how your structure, math, and communication are scored.

From Jane Street: Internships Overview, Jane Street: Quantitative Trader Internship (New York), Jane Street: Strategy and Product Internship (New York), Jane Street: Interviewing Guide and FAQs, Jane Street: Trading Interviews, Jane Street: Probability & Markets Guide (PDF), Jane Street: Strategy & Product Interviews, Jane Street: Strategy and Product, Jane Street: Preparing for a Software Engineering Interview, Jane Street: Trading Internship Curriculum, Jane Street: Departments and eFinancialCareers: What Interns at Jane Street Actually Do.

1 practice question and sample answer

Say your answer out loud first, then read the sample answer.

Practice question

  1. We are going to play a market-making game. I will roll two standard six-sided dice (2d6) in secret, and the contract settles to the sum of the two dice in dollars. Make me a market, 10 up. Then, suppose I immediately say 'Take 'em' at your offer, and tell you that at least one of the dice is a 6. How do you update your market for the next round?

    Sample answer. Example answer written by our team. The applicant is invented. First, I will find the unconditional expected value of the sum of two standard six-sided dice. By linearity of expectation, the expected sum of two dice is the sum of their individual expected values. Each d6 has an expected value of 3.5, so 3.5 plus 3.5 gives $7.00. For my initial market around $7.00, I will say, 'I'm 6.50 at 7.50, 10 up.' When you say 'Take 'em,' you buy 10 contracts at $7.50, and then reveal that at least one die is a 6. I need to update my expected value using conditional probability. Out of the 36 equally likely ordered pairs for two dice, there are 6 outcomes where the first die is 6, and 6 where the second die is 6. Subtracting the double-counted [6,6] leaves 11 equally likely outcomes. 10 of those outcomes pair a 6 with 1 through 5 (2 each of sums 7, 8, 9, 10, and 11, totaling 90) plus 1 [6,6] outcome of 12, which equals 102. Dividing 102 by 11 gives an updated expected value of 9 and 3/11, or about $9.27. Because you knew this before lifting my $7.50 offer, I faced adverse selection. Centered around $9.27, my updated market is 'I'm 8.75 at 9.75, 10 up.'

    Uses linearity of expectation to get the initial $7.00 fair value quickly, states the quote in Jane Street's exact market-making phrasing, counts all 11 conditional outcomes accurately without double-counting [6,6], and explains the adverse selection of the first fill.

Practice the interview, out loud

Soreno's AI interviewer asks banking fit and technical questions, follows up on your answers and scores them.

Questions people ask

How much do Jane Street interns get paid?

According to Jane Street's 2026 New York job postings, the annualized base salary is $300,000 for the Quantitative Trader Internship and $250,000 for the Strategy and Product Internship.

Do I need a finance background, a minimum GPA, or OCaml experience to apply?

No. Jane Street has no GPA or degree requirement, does not require prior knowledge of finance or economics, and does not require a cover letter. For coding rounds, you can use whichever programming language you know best, and Jane Street strongly recommends not trying OCaml for the first time during an interview.

Can I apply to multiple Jane Street internship roles or offices at once?

Yes. Whenever you submit an application, Jane Street automatically evaluates you for all open roles globally and helps match you to the right team, so it is common to be considered for more than one track during your interviews.

When is the deadline to apply for a Jane Street summer internship?

Jane Street reviews applications on a rolling basis with no firm-wide deadline, and recommends applying as soon as you feel ready to interview within the next few weeks.

What happens if I get an answer wrong in a Jane Street trading interview?

You do not automatically fail. Jane Street designs its interviews as collaborative conversations and evaluates how you approach unfamiliar problems, ask clarifying questions, take hints, and correct mistakes.

Can I reapply to Jane Street if I was rejected before?

Yes, many full-time employees were hired on their second or third attempt. For current students, Jane Street recommends waiting at least a year before reapplying so you have new classroom or project experiences to share.

Keep going