Jane Street Interview Process in 2026: Math, OCaml & Puzzles

Jane Street is one of the most selective technical employers on the planet. They hire fewer engineers and traders per year than most FAANG teams hire in a single quarter, and they are unapologetic about it. If you are aiming for a software engineering or quantitative trading role at Jane Street in 2026, you need a preparation strategy that is fundamentally different from the LeetCode grind you used to land your current job. The interview is less about data structures and more about how you think under pressure—and whether you can do it in OCaml. This guide tells you exactly what to expect, what actually matters, and how to spend your prep time.

The Jane Street Hiring Funnel Is Designed to Filter, Not Welcome

Jane Street does not post many open roles, does not have a giant recruiting machine, and does not particularly care if the process feels intimidating. That is a feature, not a bug. The firm's entire business model depends on hiring people who are genuinely exceptional at probabilistic reasoning and first-principles thinking. They would rather have an empty seat than a bad hire.

In 2026, the standard software engineering pipeline looks roughly like this:

1. Resume screen or recruiter outreach (Jane Street does a lot of direct sourcing)
2. An initial phone screen focused on probability and mental math
3. A technical phone interview covering OCaml and functional programming
4. A full on-site or virtual on-site: typically four rounds over one day covering algorithms, systems thinking, probability, and a market-making or trading simulation
5. A final debrief and offer call if everything aligns

Trading and quantitative research roles follow a similar path but lean even harder on probability, expected value problems, and live trading simulations. This guide focuses primarily on the software engineering track, with relevant notes for quant-adjacent roles.

Mental Math Is Not a Warm-Up — It Is a Core Signal

Every candidate who has been through Jane Street interviews says the same thing afterward: they did not expect the mental math to be so serious. Let's be direct — if you cannot quickly estimate, compute, and sanity-check numbers in your head, Jane Street is not the right destination right now. Fix that first before anything else.

The mental math questions are not trick questions. They are calibration questions. Interviewers want to see:

• Whether you can decompose a complex calculation into manageable steps
• Whether you check your own work by approaching the same problem from two angles
• Whether you stay calm when you get a wrong intermediate result and have to backtrack
• Whether you can give a confident answer while being honest about your uncertainty

Expect questions like: "If a stock moves up 10% and then down 10%, what is the net change?" (Answer: -1%, not zero — and you should get this in under three seconds.) Or: "Estimate the number of seconds in a year to two significant figures." These are not hard if you practice. They feel hard under pressure if you have never drilled them.

The best preparation resource for this remains Secrets of Mental Math by Arthur Benjamin, combined with daily timed drills on sites like mathtrainer.org. Spend at least four weeks on mental math before your first screen. Thirty minutes a day. Non-negotiable.

OCaml Is Not Optional — Learn It Properly

Jane Street is the most prominent real-world user of OCaml in the industry. Their internal codebase is almost entirely OCaml, and they expect software engineering candidates to code in it during interviews. If you show up planning to use Python or Java, you will not get an offer.

The good news: OCaml is not that hard to learn at the surface level needed for interviews. The bad news: functional programming requires a genuine mental model shift, and you cannot fake it. Interviewers at Jane Street have been writing OCaml professionally for years. They will immediately tell if you are translating imperative code into OCaml syntax versus actually thinking functionally.

Here is what you need to be fluent in before an interview:

• Pattern matching on variants, lists, and tuples
• Recursive function design (especially tail-recursive variants)
• Higher-order functions: map, fold_left, fold_right, filter
• The Option type and how to handle None cases idiomatically
• Basic module system and how to structure a small program
• Immutable data by default — thinking in terms of returning new values, not mutating

You do not need to know Jane Street's internal libraries (Jane, Core, Async) for interviews, but knowing they exist and understanding why Jane Street built them demonstrates genuine interest. The canonical learning path in 2026 is: Real World OCaml (freely available online, second edition updated for OCaml 5) → OCaml Exercises on exercism.io → solving at least ten LeetCode-style problems in OCaml from scratch before your interview.

"The candidates who impress us most are not the ones who know OCaml trivia. They are the ones who write idiomatic functional code naturally, like it is how they actually think." — paraphrased from multiple Jane Street engineering blog posts and public talks

Give yourself six to eight weeks of regular OCaml practice. Do not treat it as a checkbox. Treat it as learning a new way to think about programs.

Probability and Puzzle Rounds Are Where Offers Are Won or Lost

This is the heart of the Jane Street interview for both engineering and trading tracks. The probability and puzzle questions test your ability to reason from first principles, build models quickly, and communicate your reasoning clearly while you are still working through the problem.

Common question archetypes include:

• Expected value calculations: "You flip a coin. Heads you win $2, tails you lose $1. What is this game worth? Now what if you can choose to stop after any flip?"
• Geometric probability: Problems involving random points, random chords, random intersections
• Bayesian reasoning: Conditional probability updates, often framed as market scenarios
• Game theory and optimal strategy: Multi-round games where you must choose a strategy, not just compute a single answer
• Fermi estimation under constraints: Back-of-envelope calculations that require you to be explicit about your assumptions

The key insight most candidates miss: Jane Street interviewers do not just want the right answer. They want to see you build a model out loud, check it, and update it when they give you new information. They will often deliberately give you problems with no clean closed-form answer to see how you handle ambiguity.

The best preparation materials for this are Heard on the Street by Timothy Crack, Fifty Challenging Problems in Probability by Frederick Mosteller, and the Jane Street puzzle archive on their website (they publish monthly puzzles). Work through at least twenty to thirty probability problems out loud, with someone listening, before your interview. Solo studying is insufficient here — you need to practice verbalizing your reasoning.

The Coding Rounds Emphasize Correctness and Clarity Over Speed

Jane Street's coding interviews are not LeetCode sprints. They are closer to pair programming sessions where the interviewer genuinely wants to see how you approach a problem from zero. The expectations:

• Write code that is correct before trying to make it clever
• Handle edge cases proactively and explain why you are handling them
• Use functional patterns naturally — recursion over loops, immutability by default
• Be willing to step back and redesign if your initial approach is suboptimal
• Ask clarifying questions before writing a single line

Typical problem difficulty is roughly LeetCode medium to hard, but the framing is more abstract. You might be asked to implement a simplified order book, build a small interpreter, or write a function that models some probabilistic process. The problems are designed so that an elegant functional solution is noticeably cleaner than a brute-force imperative one.

One underrated piece of advice: practice explaining why you chose a particular data structure or algorithm. Jane Street engineers think carefully about tradeoffs, and they want to see that you do too. Saying "I'm using a map here because lookup is O(log n) and insertion order doesn't matter" is better than just writing the code silently.

The Trading Simulation Round Is Real and You Should Prepare for It

For software engineering candidates interviewing for roles that touch trading infrastructure, and for all quant trading candidates, there is typically a live trading simulation round. You will be placed in a mock market and asked to make bids and offers on some asset or game. This round tests:

• Whether you can form and update probability estimates in real time
• Whether you manage risk sensibly (not just maximize expected value greedily)
• Whether you can handle adversarial counterparties — the interviewers are deliberately trying to trade against you
• Whether you stay composed when you take losses

The most important thing to understand about this round is that being too passive is as penalized as being reckless. If you refuse to make markets because you are uncertain, that is a red flag. Uncertainty is the entire context. You are supposed to be making decisions under uncertainty.

Prepare by reading Trading and Exchanges by Larry Harris for market microstructure intuition, and by playing Jane Street's public trading games (they have released several, including Figgie, their custom card game designed to simulate market dynamics). Take Figgie seriously — it is not a novelty, it is a teaching tool built by traders.

Compensation at Jane Street in 2026 Is Exceptional — and So Are the Expectations

Let's talk numbers, because this is presumably part of why you are considering Jane Street in the first place.

For software engineering roles in 2026, total compensation at Jane Street for new graduate hires is in the range of $250,000–$350,000 USD all-in, depending on level. Experienced hires at senior and principal equivalent levels routinely see $400,000–$600,000+ including bonus. For quantitative trading roles, the upside is significantly higher because compensation is more directly tied to trading performance, with top performers earning well into seven figures.

For Canadian-based candidates like Alex Chen: Jane Street's primary offices are New York, London, and Hong Kong. They do not have a Vancouver presence, and remote-only arrangements are essentially unheard of for engineering roles at Jane Street given the firm's culture of tight collaboration and information security requirements. This is a real constraint worth naming honestly — if relocation is truly off the table, Jane Street is likely not a fit regardless of technical ability.

The compensation is exceptional because the job is genuinely hard, the firm has an extremely high revenue-per-employee ratio, and they want people who are so good that paying them generously is obviously the right business decision.

Next Steps

If you are serious about Jane Street in 2026, here is what you should do in the next seven days:

1. Assess your OCaml level honestly. Spend two hours trying to solve three medium-difficulty algorithm problems in OCaml from scratch. If you struggle to write a clean recursive list manipulation without Googling, you need six to eight weeks of dedicated practice before you are ready to apply.
2. Start daily mental math drills today. Twenty to thirty minutes of timed arithmetic and estimation practice. Use mathtrainer.org and work through the first four chapters of Secrets of Mental Math. Set a reminder. Do it every day without exception.
3. Solve one Jane Street monthly puzzle. Go to janestreet.com/puzzles and pick a recent one. Work through it completely, write up your solution, and then read the official solution. Notice where your approach diverged.
4. *Get a copy of Heard on the Street and work through ten probability problems out loud.* Record yourself if you do not have a study partner. Listen back. Notice where your verbal reasoning gets sloppy or circular.
5. Play Figgie with real people. Download the rules, get three other people, and play several rounds. Pay attention to how your probability estimates change as cards are revealed, and whether you are making markets or just reacting to other people's markets.

Jane Street is hard to get into by design. But the process is genuinely fair — they are testing real skills that they actually use, not arbitrary hazing rituals. If you can do the math, write clean functional code, and reason clearly under pressure, you have a real shot. Start with step one above, and be honest with yourself about how far you have to go.

Sources and further reading

When evaluating any company's interview process, hiring bar, or compensation, cross-reference what you read here against multiple primary sources before making decisions.

• Levels.fyi — Crowdsourced compensation data with real recent offers across tech employers
• Glassdoor — Self-reported interviews, salaries, and employee reviews searchable by company
• Blind by Teamblind — Anonymous discussions about specific companies, often the freshest signal on layoffs, comp, culture, and team-level reputation
• LinkedIn People Search — Find current employees by company, role, and location for warm-network outreach and informational interviews

These are starting points, not the last word. Combine multiple sources, weight recent data over older, and treat anonymous reports as signal that needs corroboration.

Related guides

• Adobe Interview Process in 2026 — Creative Cloud Engineering, ML, and Craft — Adobe interviews in 2026 blend practical engineering, product taste, and craft: expect coding, system design, and a lot of discussion about shipping durable tools for creative and document workflows.
• Airbnb Interview Process 2026: Craft, Values & Core Values Round — A no-fluff breakdown of Airbnb's 2026 interview process, including the craft round, core values interview, and how to actually prepare.
• The Amazon Interview Process in 2026: Loops, Bar Raisers & the Written Round — A no-fluff breakdown of Amazon's 2026 interview loop, bar raiser role, and written assessment—with actionable prep advice for SWE candidates.
• Anduril Data Scientist Interview Process in 2026 — SQL, Modeling, Experimentation, and Product Analytics Rounds — Anduril data scientist interviews in 2026 focus on SQL, modeling, experimentation, and product analytics in defense-tech systems where data is messy, high-stakes, and operational. The strongest candidates connect analysis to operator decisions, sensor reliability, field deployment, and model evaluation.
• Anduril Product Manager Interview Process in 2026 — Product Sense, Execution, Strategy, and Behavioral Rounds — Anduril PM interviews in 2026 test whether you can turn mission needs, operator workflows, hardware constraints, and defense buying dynamics into shippable products. Prepare for product sense, execution, strategy, and behavioral rounds that punish generic SaaS answers.