Four candidates describe the Jane Street Data Scientist interview as a mix of probability puzzles, unconventional live coding and open-ended data rounds. All four report no offer. Here is what each reported and how to prepare.
$226K
Avg. Base Comp
$292K
Avg. Total Comp
2-6 rounds
Typical Rounds
Not reported
Process Length
Candidates who interviewed for Jane Street's Data Scientist role describe a process that is more mathematical and open-ended than a typical data science loop. All four reports ended in no offer, and the structure they describe is not identical, so treat the stage order as a rough guide.
Probability comes up repeatedly. One candidate opened with a phone screen of coin-flip and conditional probability questions. Another faced a dice game with up to three rolls and had to find the expected gain by backward induction. The same candidate said the third online round was where difficulty jumped, with harder probability and distribution questions. Expected value and conditional probability deserve the most drilling.
Coding is reported as live and unconventional. Candidates describe design-focused problems, a market and stock-related problem, and a sorting and math puzzle about two cyclists, rather than standard LeetCode questions. Thinking aloud with the interviewer was part of it.
The data rounds are the other distinctive piece. Candidates report building a model on a trail-difficulty dataset with a holdout set, and a case study that produced a written report where presentation and detail mattered. One also faced an unstructured scenario with little context. In the onsite data round, one candidate was asked how to predict Amazon's online-only revenue from credit card transaction data, which involved survivorship bias and missing data. State your assumptions early and ask for confirmation.
One candidate described a five-round onsite day that included a behavioral session where they presented a topic of interest. They reported fatigue by the last brain teaser, so pace yourself.
Synthesized from 4 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Jane Street process.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Jane Street
Given two sorted lists, write a function to merge them into one sorted list.
| Question | |
|---|---|
| 500 Cards | |
| Cumulative Sales Since Last Restocking | |
| Total Transactions | |
| Ride Coupon | |
| Stranded Miner | |
| Poker Pair | |
| Priority Queue Using Linked List | |
| Assumptions of Linear Regression | |
| Stick Break | |
| Simulating Coin Tosses | |
| Second Ace | |
| Estimating D | |
| Last Element of a Singly Linked List | |
| Above Average Product Prices | |
| International e-Commerce Warehouse | |
| Summing Numeric Strings | |
| Text Editor With OOP | |
| Identifying Good Investors | |
| Incentive Scheme | |
| LRU Cache 1 | |
| Credit Score Estimation | |
| Decreasing Ride Costs | |
| WallStreetBets Sentiment Analysis | |
| 2nd Highest Salary | |
| Empty Neighborhoods | |
| Rolling Bank Transactions | |
| Employee Salaries | |
| Subscription Overlap | |
| Comments Histogram |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report the process opening with probability-heavy questions. One described a 45-minute phone screen of coin-flip and conditional probability questions. Another described an online round of brain teasers, including a three-roll dice game solved by backward induction. Expect expected value and conditional probability, not coding.
Candidates describe coding with an engineer or team member in real time. Problems were reportedly not standard LeetCode: one was design-focused and easier, one was a longer market and stock-related problem, and one was a sorting and speed-and-distance puzzle. Thinking aloud appears to matter, and candidates may work in a language of their choice.
One candidate says the third online round had much harder probability and distribution questions. Others report data-oriented rounds: building a model for trail difficulty on a holdout set, or producing a written case report. One candidate mentions out-of-sample performance and polish on the written work. Not every candidate reports this stage in the same position.
Candidates who reached the onsite report a data round where you explore a dataset with the interviewer and discuss modeling. One example was predicting Amazon's online-only revenue from credit card transaction data, where biases and missing data came up. Another reported an unstructured scenario with limited clarification, so stating assumptions explicitly is likely wise.
Onsite formats varied between candidates. One described two morning interviews mixing math and machine learning, then being sent home after lunch. Another described five rounds across a day: coding, data analysis, a behavioral presentation of a topic of interest, and two brain teasers. Fatigue was mentioned, so you may want to practice switching between formats.
