
Company interview guide
JPMorgan Chase Machine Learning Interview Guide (2026)
A recorded first-round assessment, then team interviews that test ML basics, your own projects and, for quant tracks, probability and statistics.
JPMorgan Chase's machine learning interview loop has 6 reported stages (HR or recruiter screen, Virtual assessment: recorded video and online coding test, Technical screen: Python and data, ML fundamentals and resume deep dive, Probability, statistics and math (quant and data science tracks), Team interviews and fit); typical length: Varies by team; one report spread three rounds over about six weeks.
Roles covered: ML Engineer, AI Engineer, Data Scientist and Research Scientist. The question bank holds 13 questions reported from JPMorgan Chase interviews, 1 of them free to solve in your browser.
- Interview process length
- Varies by team; one report spread three rounds over about six weeks
- Stages
- 6
- Reports linked to bank questions
- 13
- Bank report dates
- Feb 2022–Jun 2026
- Questions in bank
- 13 (1 free)
- Last reviewed
- October 2026
How the JPMorgan Chase interview loop works
An application review and, in most reports, a short HR or recruiter screen about your background. The bank's guide describes a first-round virtual assessment: a recorded video interview with set questions, plus an online coding test for technology roles. Then one or more technical rounds, usually Python or data work and ML concept questions mixed with a walk-through of your past projects, and finally several one-to-one or panel interviews with team members, held by video or in person. Quant research and some data science loops lean on probability and statistics instead of ML coding. Applications are reviewed on a rolling basis and the bank says it cannot speed up decisions. Only seven first-hand reports back this outline, and none gives a full loop end to end.
The JPMorgan Chase interview process, stage by stage
01HR or recruiter screen
Short phone or video call; length not reported
A conversation about what you do now, your experience and your field of study. Internship candidates in particular describe it as a plain background screen with a follow-up call. Use it to learn which team is hiring and how the later rounds will run.
What it tests
- A clear summary of your current work and experience
- Your field of study and how it fits the role
- Basic logistics and motivation for the team
How to prepare
- Ask which team owns the role and what each later round covers; formats differ between groups.
- Have a two-minute summary of your most relevant ML or data project ready.
An HR or screening call opens the loop in 2 ML-role reports from 2024 and 2025; other reports do not say how the process started.
02Virtual assessment: recorded video and online coding test
Recorded video interview with set questions and a time window; technology roles may add an online coding test
The bank's candidate handbook names a recorded video interview and, for technology roles, an online coding test as the usual first round, and its FAQ adds that many assessments cannot be retaken for a set period. None of our first-hand ML reports describes this round, so it may be skipped for experienced hires. Compiled campus reports for quant research describe a timed coding test of two algorithm questions in about an hour, then a short recorded interview with brief prep time per question.
What it tests
- Short behavioral answers recorded against a timer
- Standard algorithm coding in a single function
- Why the bank and why this division or team
How to prepare
- Answer recorded questions in a situation, action, result shape and finish inside the time limit.
- Prepare one story each for teamwork, a mistake, persistence and explaining something complex.
- Treat the coding test as a one-shot attempt; the bank says many assessments cannot be retaken for a while.
The company's handbook and hiring FAQ describe a recorded video interview and an online coding test for technology roles; no first-hand ML report in our set covers it.
03Technical screen: Python and data
Live video round with a team member; length not reported
Reported screens are practical rather than puzzle-heavy. One senior data science candidate did exploratory analysis in NumPy on a given dataset, finding, treating and plotting outliers, plus a Python question. A 2026 ML engineer screen opened with questions on the candidate's own AI work and ended with basic Python coding: return every index pair summing to a target, with repeated follow-ups on hash-map design and complexity.
What it tests
- Exploratory data analysis and outlier handling in NumPy
- Basic Python coding with hash maps and duplicate values
- Explaining time complexity and trade-offs against brute force
- Talking through your recent ML or AI work before the coding
How to prepare
- Be ready to justify every data-structure choice; one interviewer kept asking why a map stores a list of indices.
- Practise a quick EDA routine: summary stats, an outlier rule, a fix and a plot.
- Expect easy to medium difficulty, judged on clear reasoning.
Two ML or data science reports from 2025 and 2026 describe a first technical round: one on NumPy EDA plus Python, one on resume questions plus hash-map coding.
04ML fundamentals and resume deep dive
One or more video rounds with team members; often mixed with project discussion
Candidates describe direct knowledge checks on core ML: what machine learning is and how a model moves from development to production, detailed SVM questions, tree methods, learning approaches and tuning. Interviewers also probe what you personally did; a 2026 screen for an AI role asked which models an agent workflow used, how the orchestrator was built, and who owned endpoints, data access and entitlements.
What it tests
- Classical ML in depth: SVMs, decision trees and ensembles, tuning
- The ML development lifecycle, from data to a deployed model
- Your own contribution to each project on your resume
- For AI roles, agent and orchestration design choices
How to prepare
- Revise SVM margins, kernels and hinge loss plus tree ensembles and tuning; an intern got SVM and tree questions, a senior associate got tuning.
- For every resume line, know what you built yourself and what the team did.
- For AI roles, be ready to explain your multi-agent or tool-use setup concretely.
ML concept questions or a project deep dive in 4 ML-role reports from 2022 to 2026; content is consistent (classical ML basics, tuning, own projects) but details are thin.
05Probability, statistics and math (quant and data science tracks)
Questions inside technical rounds; brain teasers can appear in ML rounds too
One ML intern got brain teasers next to the ML questions. Beyond that, the support is compiled question lists from quant research and data science candidates, which lean on dice and conditional-probability puzzles, OLS derivations and tests, hypothesis-testing trade-offs, missing data and, for quant roles, proofs from linear algebra and numerical methods. Desk-facing quant roles reportedly add derivatives pricing and stochastic calculus.
What it tests
- Conditional probability and expected-value puzzles
- Deriving the OLS estimator and testing coefficients
- Type I and II error, power and R-squared reasoning
- Matrix and numerical-method proofs for quant research
How to prepare
- Set up the recursion or symmetry argument out loud before computing.
- Be able to derive OLS and its variance on paper and say which assumption breaks first in real data.
- If you apply to quant research, confirm with the recruiter whether pricing and stochastic calculus are in scope.
Free reported questions for this stage
- Correlation Matrices Are Positive SemidefiniteLinear Algebra & OptimizationMediumSolve
One 2023 ML intern report mentions brain teasers; the rest rests on compiled question lists from quant research and data science candidates, not first-hand loop reports.
06Team interviews and fit
Several one-to-one or panel interviews by video or in person, sometimes spread over weeks
The bank says candidates meet several people in multiple rounds, held virtually or in person depending on the program. One senior associate had three video interviews with team members over about 45 days covering ML concepts, past experience and real scenarios; another candidate faced a panel of one-to-one calls that dwelt on the resume. Fit questions reported include why you are changing jobs and whether you would take a role closer to back or middle office.
What it tests
- Applying ML to real business scenarios
- Depth and honesty about past experience
- Motivation for the bank and the specific team
- Openness to where the role sits in the firm
How to prepare
- Prepare two questions about the team and program, as the bank's own guide advises.
- Have a clear answer on why you are moving and what kind of team placement you would accept.
- Do not overstate your role; a few reports describe close questioning of resume claims.
Two ML-role reports (2022, 2025) describe multiple one-to-one team interviews; the company's hiring page says candidates meet several people in multiple rounds.
Questions and themes reported most often
- Defend which parts of your resume projects you did yourself (3 reports, ml fundamentals and resume deep dive)
- Explain core ML: SVMs, tree methods, ensembles and tuning (2 reports, ml fundamentals and resume deep dive)
- What machine learning is and how a model goes from development to production (1 reports, ml fundamentals and resume deep dive)
- Return every index pair summing to a target, handling duplicates with a hash map (1 reports, technical screen: python and data)
- EDA on a dataset: find, treat and plot outliers in NumPy (1 reports, technical screen: python and data)
- Agent workflow deep dive: models, orchestration, tool use and data entitlements (1 reports, ml fundamentals and resume deep dive)
- Brain teasers alongside ML questions (1 reports, probability, statistics and math (quant and data science tracks))
- Why are you changing jobs, and would you take a back or middle-office role (1 reports, team interviews and fit)
What JPMorgan Chase asks about
- ML fundamentals & project deep dive30%
- Probability & statistics30%
- Python & data coding20%
- Behavioral & fit20%
JPMorgan Chase interview prep checklist
- Can explain SVMs, tree ensembles and hyperparameter tuning without notes
- Solved an all-pairs-sum problem with duplicates and can defend its complexity
- Ran an EDA on a fresh dataset: outliers found, treated and plotted
- Know my personal contribution to every project on my resume
- Derived OLS and worked a few dice and conditional-probability puzzles (quant and data science tracks)
- Recorded practice answers to timed behavioral questions
- Asked the recruiter for the type and length of each round
Tips
- Ask the recruiter early which team owns the role and what each round covers; loops differ between groups.
- Revise classical ML thoroughly (SVMs, trees, boosting, tuning); it is the most consistent technical theme in reports.
- Prepare to say exactly what you built in every project, including data access and deployment details.
- Drill easy to medium Python with hash maps and be ready for several follow-ups on complexity and edge cases.
- Practise a short NumPy or pandas EDA routine with outlier handling and a plot.
- For quant or data science tracks, work through probability puzzles and OLS and hypothesis-testing derivations on paper.
- Rehearse timed recorded answers in a situation, action, result shape for the first-round video interview.
- Expect gaps of weeks between rounds and plan your other applications around it.
See the rest of the JPMorgan Chase loop
The later stages, the questions reported for them and how to prepare are in Pro.
Unlock with ProPractice the JPMorgan Chase loop
Work the stages in order: every reported question for this company, grouped by round, with solutions in the app.
JPMorgan Chase interview FAQ
How many interview rounds does JPMorgan Chase have for ML roles?
Candidates report 6 stages: HR or recruiter screen, Virtual assessment: recorded video and online coding test, Technical screen: Python and data, ML fundamentals and resume deep dive, Probability, statistics and math (quant and data science tracks) and Team interviews and fit. An application review and, in most reports, a short HR or recruiter screen about your background. The bank's guide describes a first-round virtual assessment: a recorded video interview with set questions, plus an online coding test for technology roles. Then one or more technical rounds, usually Python or data work and ML concept questions mixed with a walk-through of your past projects, and finally several one-to-one or panel interviews with team members, held by video or in person. Quant research and some data science loops lean on probability and statistics instead of ML coding. Applications are reviewed on a rolling basis and the bank says it cannot speed up decisions. Only seven first-hand reports back this outline, and none gives a full loop end to end.
How long does the JPMorgan Chase ML interview process take?
Reported timelines run Varies by team; one report spread three rounds over about six weeks.
What does JPMorgan Chase ask in ML interviews?
By reported question mix: ML fundamentals & project deep dive 30%; Probability & statistics 30%; Python & data coding 20%; Behavioral & fit 20%. The most repeated themes: Defend which parts of your resume projects you did yourself; Explain core ML: SVMs, tree methods, ensembles and tuning; What machine learning is and how a model goes from development to production.
How should I prepare for the JPMorgan Chase ML interview?
Ask the recruiter early which team owns the role and what each round covers; loops differ between groups. Revise classical ML thoroughly (SVMs, trees, boosting, tuning); it is the most consistent technical theme in reports. Prepare to say exactly what you built in every project, including data access and deployment details.
Are there free JPMorgan Chase interview questions?
Yes. LeetML has 13 questions reported from JPMorgan Chase interviews, and 1 of them are free to solve in the browser; they are listed by stage above.
Related interview guides
- CitiA thinly documented loop: an application and skills check, a background interview, then a final stage of technical interviews, with quant roles leaning on probability and markets maths.
- Bank of AmericaA conventional bank loop: recruiter and hiring-manager calls, a technical conversation, sometimes a take-home, with statistics-heavy questions for quant roles.
- Capital OneA timed four-question coding assessment, then a four-round Power Day of ML technical, a business case, behavioral and, on recent ML engineer loops, ML system design.
- BloombergClassic coding rounds, solid ML and NLP fundamentals and an open-ended model design, usually with two interviewers per round.