Company interview guide
Walmart Machine Learning Interview Guide (2026)
Wide ML and statistics questioning, easy-to-medium coding and retail-flavoured ML design, run team by team.
Walmart's machine learning interview loop has 8 reported stages, from recruiter or hr call to hiring manager and behavioral; typical length: 3 weeks to several months.
Roles covered: Data Scientist and ML Engineer. The question bank holds 8 questions reported from Walmart interviews, 1 of them free to solve in your browser.
- Interview process length
- 3 weeks to several months
- Stages
- 8
- Reports linked to bank questions
- 9
- Bank report dates
- Dec 2023–Oct 2026
- Questions in bank
- 8 (1 free)
- Last reviewed
- October 2026
How the Walmart interview loop works
Usually a recruiter or HR contact, often an online assessment, then a first technical round and a set of about three to five virtual interviews of roughly an hour each, ending or starting with the hiring manager. Several candidates say each round was booked only after good feedback on the previous one, so the loop can stretch over weeks or months. India-based loops since 2024 tend to name the rounds ML breadth, ML depth, ML coding, ML design and hiring manager; US data scientist loops more often mix Python coding, SQL, a case study and a data science technical round. A few teams use a 24-hour take-home with a presentation instead of live rounds. Reports since 2024 are fewer than older ones, so treat round counts as typical rather than fixed.
The Walmart interview process, stage by stage
01Recruiter or HR call
Short phone call, about 20-30 min; some candidates only get an email form
A call about your background, salary expectations, relocation and why Walmart. In several recent reports this step was thin or skipped: the recruiter sent a written form, or the candidate went straight to a hiring-manager call. A few candidates were contacted and then not followed up, so confirm next steps in writing.
What it tests
- Why Walmart and why this team
- Salary expectations, location and relocation
- A short summary of relevant data science or ML work
How to prepare
- Ask which team you are interviewing for; content varies a lot by team (supply chain, ads, catalog, search).
- Ask how many rounds there are and whether each one depends on the last, so you can plan your timeline.
- Follow up by email if you hear nothing within a week or two.
A recruiter or HR call is described in 11 data scientist and intern reports, 2021-2025; only 3 are from 2024 or later, and some recent candidates got an email form or went straight to the hiring manager.
02Online assessment
Self-paced online test: coding plus SQL, probability and ML multiple choice; some teams send a coding-only screen
Many data science loops start with a timed online test you take on your own time. Reports describe one or two coding questions, multiple-choice questions on SQL, Python and basic probability, and in older reports a time series modelling task and a behavioral questionnaire. A 2026 senior data scientist screen was coding only, with a recursion problem. Some interviewers later walk through your assessment answers.
What it tests
- Easy-to-medium coding in Python
- SQL joins, aggregation and window functions
- Probability and ML basics in multiple-choice form
- Fitting a simple time series model under time pressure
How to prepare
- Be ready to explain every assessment answer; one 2026 interview opened with a review of the candidate's solution.
- Revise SQL window functions and basic probability as well as coding; the multiple-choice part is broad.
- Write clean, commented code since reviewers read it later.
An online assessment appears in 12 data scientist, ML and intern reports, 6 of them since 2024; contents vary from coding only to coding plus SQL, probability and ML multiple choice.
03First technical round
About 45-90 min video call with a senior data scientist, the hiring manager or a small panel; some ML roles used an outside interviewing firm
The first live round usually mixes three things: a walk-through of one or two projects on your resume with a focus on design choices, quick ML and statistics questions, and one or two easy-to-medium coding problems at the end. Some teams make it a statistics and probability round instead, and one 2026 supply chain screen had three interviewers covering SQL, Python, probability and time series in 90 minutes. Several candidates say interviewers kept their camera off.
What it tests
- Explaining design choices in your own projects
- Core ML questions: overfitting, regularization, tree models, metrics
- Statistics and probability fundamentals
- One or two easy-to-medium array or string coding problems
How to prepare
- Prepare a crisp deep dive on the two projects closest to the team's work, including why you chose each model.
- Leave time for coding: it often comes in the last 20-30 minutes after a long discussion.
- If the team is in supply chain or forecasting, revise time series methods and evaluation.
A first technical round of projects, ML or stats questions and short coding is described in 21 data scientist and ML reports, 10 of them since 2024; its interviewer and emphasis vary by team.
04Take-home and presentation (some teams)
Usually a 24-hour take-home, then a presentation or a detailed discussion of your solution
Some data scientist teams replace or add to live rounds with a take-home, typically a binary classification or modelling task due within a day, followed by a presentation to the team. Interviewers question the algorithms and data handling you used in detail. This format is mostly in older reports; only one report since 2024 describes it.
What it tests
- Building a sound baseline model quickly
- Justifying feature, model and metric choices
- Presenting results clearly to a team
How to prepare
- Keep the solution simple and well explained rather than heavily tuned.
- Know every method and library call you used; follow-up questions go deep.
- Prepare a short slide or notebook summary with results and limitations.
A take-home, usually with a presentation or walk-through, appears in 5 data scientist reports from 2018-2024; only one is from 2024, so most current loops seem to use live rounds instead.
05Coding rounds
Live coding in Python, often combined with ML questions in the same hour; SQL in many data scientist loops
Coding shows up in most loops, either as its own round or as the last part of a technical round. Problems are mostly easy to medium: arrays, strings, sorting, dynamic programming, linked lists and grids. Many teams also ask you to implement an ML or probability routine without libraries, such as sampling from a discrete distribution, simulating a biased die, kNN, logistic regression or a small neural network, and data scientist loops often add SQL or PySpark data manipulation.
What it tests
- Easy-to-medium algorithm problems with follow-ups
- Implementing ML algorithms from scratch without numpy or sklearn
- Sampling and simulation from a given distribution
- SQL joins and window functions, sometimes PySpark
How to prepare
- Practise writing kNN, logistic regression and a sampler in plain Python, not just with libraries.
- Do a set of medium array, sort and dynamic programming problems; reports rarely mention hard ones.
- Revise SQL window functions alongside Python.
Live coding is described in 36 data scientist and ML reports, 20 of them since 2024; most call the problems easy to medium, and at least 6 asked for an ML or sampling routine from scratch.
06ML breadth, depth and statistics
One or more rounds of about an hour; India loops often split ML breadth and ML depth
The most consistent part of the loop. Expect many quick questions on classic ML: random forest versus gradient boosting, tuning a GBM, logistic and linear regression, regularization, clustering, Naive Bayes and overfitting, plus statistics such as regression assumptions, hypothesis tests and type I and II errors. Depth rounds pick one area from your resume or the team's work, such as time series forecasting, NLP, contrastive learning or transformers, and push on it. Some candidates found the questions textbook style rather than practical.
What it tests
- Tree ensembles, regression, regularization and clustering
- Statistics: hypothesis tests, error types, regression assumptions
- Time series forecasting for supply chain and demand teams
- Deep learning and transformer basics, including attention
How to prepare
- Drill short, exact definitions: candidates report a rapid series of textbook questions.
- Know why L1 regularization zeroes weights and how random forests and boosting differ in bias and variance.
- Prepare one depth area that matches the team, and be ready to derive or write the key formula.
- Expect some generative AI and transformer questions in 2025-2026 loops.
Free reported questions for this stage
- Random Forest vs Gradient Boosting (XGBoost)Classical MLMediumSolve
ML and statistics questions appear in 42 data scientist and ML reports from 2018-2026, the most consistent stage; India reports since 2021 often name separate ML breadth and ML depth rounds.
07ML system design or case study
About 60 min, sometimes running long; staff-level loops can have several design rounds
Senior and staff loops, and many India loops, include an ML design round built on retail problems: detecting catalog listings that break policy, summarizing customer reviews, an ad targeting system, recommendations or an agentic product search. US data scientist loops more often frame it as a case study where you clarify the data, propose a solution and then improve it. One staff candidate had four design rounds and a director round with no coding at all.
What it tests
- Framing a retail or marketplace problem as an ML task
- Data, labels, features and model choice at catalog scale
- Evaluation and monitoring, including drift after launch
- When to use an LLM or agent versus a simpler model
How to prepare
- Ask clarifying questions about the data before proposing a model; interviewers reward it.
- Prepare one catalog or marketplace design (policy violations or fraud) and one LLM design (review summaries or agentic search).
- Cover how you would measure success offline and online, and how you would handle a wrong answer.
An ML design or case round appears in 14 data scientist and ML reports, 10 of them since 2024; design prompts are retail and catalog problems, while US data scientist loops lean to case studies.
08Hiring manager and behavioral
About 30-60 min with the hiring manager, a director or HR; may come first or last
Almost every loop includes a hiring-manager conversation, but its place and content vary. Some managers open the process with a technical screen of ML and statistics; others close it with a fit round about your past work, business impact and how you collaborate. A 2026 ML engineer onsite asked about resolving a technical conflict with another team. Walmart's own careers pages advise preparing answers in the STAR format.
What it tests
- Business impact and KPIs of your past projects
- Collaboration and conflict with other teams
- Communication with non-technical partners
- Fit with the team's domain and the job description
How to prepare
- Prepare two or three STAR stories, including one about a disagreement with another team.
- Quantify the business impact of each project you mention.
- Expect technical questions even in this round; several managers asked ML fundamentals.
A hiring-manager, director or behavioral round appears in 26 data scientist and ML reports, placed first in some loops and last in others. The company's careers pages advise preparing STAR-format answers.
Questions and themes reported most often
- Random forest versus gradient boosting, and how to tune a GBM (7 reports, ml breadth, depth and statistics)
- Time series forecasting methods and evaluation (7 reports, ml breadth, depth and statistics)
- Statistics basics: hypothesis tests, type I and II errors, regression assumptions (7 reports, ml breadth, depth and statistics)
- Overfitting, regularization and why L1 gives sparse weights (5 reports, ml breadth, depth and statistics)
- Implement an ML algorithm from scratch: kNN, logistic regression, GBM or a small neural network (4 reports, coding rounds)
- Transformers, attention and LLM basics (4 reports, ml breadth, depth and statistics)
- Retail ML design: policy-violating listings, review summaries, ad targeting, agentic search (3 reports, ml system design or case study)
- Naive Bayes and probability puzzles (3 reports, ml breadth, depth and statistics)
What Walmart asks about
- ML & statistics concepts30%
- General coding22%
- ML system design & case15%
- Behavioral & hiring manager12%
- SQL & data manipulation11%
- Project deep dive10%
Walmart interview prep checklist
- Can explain random forest versus boosting, L1 versus L2 and logistic regression without notes
- Wrote kNN, logistic regression and a discrete-distribution sampler in plain Python
- Solved a set of medium array, sort and dynamic programming problems plus SQL window queries
- Revised hypothesis testing, error types and regression assumptions
- Prepared two project deep dives with design choices and measured impact
- Sketched a policy-violation or review-summarization design with metrics and monitoring
- Have STAR stories ready, including a cross-team disagreement
Tips
- Ask early which team you are interviewing with; supply chain, ads, catalog and search teams emphasize different topics.
- Make classic ML your first priority: tree ensembles, regression, regularization, clustering and Naive Bayes come up in most reports.
- Revise statistics and probability; several rounds were mostly hypothesis tests and probability puzzles.
- Practise implementing small ML algorithms and samplers in plain Python without numpy.
- Keep coding practice at easy-to-medium arrays, sorting and dynamic programming, and add SQL window functions.
- Prepare a deep dive on two projects with design choices, trade-offs and business impact.
- For senior roles, rehearse one catalog-scale ML design and one LLM or agent design end to end.
- Expect gaps of weeks between rounds and keep other processes active until you have an offer.
See the rest of the Walmart loop
The later stages, the questions reported for them and how to prepare are in Pro.
Unlock with ProPractice the Walmart loop
Work the stages in order: every reported question for this company, grouped by round, with solutions in the app.
Walmart interview FAQ
How many interview rounds does Walmart have for ML roles?
Candidates report 8 stages: Recruiter or HR call, Online assessment, First technical round, Take-home and presentation (some teams), Coding rounds, ML breadth, depth and statistics, ML system design or case study and Hiring manager and behavioral. Usually a recruiter or HR contact, often an online assessment, then a first technical round and a set of about three to five virtual interviews of roughly an hour each, ending or starting with the hiring manager. Several candidates say each round was booked only after good feedback on the previous one, so the loop can stretch over weeks or months. India-based loops since 2024 tend to name the rounds ML breadth, ML depth, ML coding, ML design and hiring manager; US data scientist loops more often mix Python coding, SQL, a case study and a data science technical round. A few teams use a 24-hour take-home with a presentation instead of live rounds. Reports since 2024 are fewer than older ones, so treat round counts as typical rather than fixed.
How long does the Walmart ML interview process take?
Reported timelines run 3 weeks to several months.
What does Walmart ask in ML interviews?
By reported question mix: ML & statistics concepts 30%; General coding 22%; ML system design & case 15%; Behavioral & hiring manager 12%; SQL & data manipulation 11%; Project deep dive 10%. The most repeated themes: Random forest versus gradient boosting, and how to tune a GBM; Time series forecasting methods and evaluation; Statistics basics: hypothesis tests, type I and II errors, regression assumptions.
How should I prepare for the Walmart ML interview?
Ask early which team you are interviewing with; supply chain, ads, catalog and search teams emphasize different topics. Make classic ML your first priority: tree ensembles, regression, regularization, clustering and Naive Bayes come up in most reports. Revise statistics and probability; several rounds were mostly hypothesis tests and probability puzzles.
Are there free Walmart interview questions?
Yes. LeetML has 8 questions reported from Walmart interviews, and 1 of them are free to solve in the browser; they are listed by stage above.
Related interview guides
- AmazonLeadership Principle questions in every round, standard algorithm coding, and ML breadth, depth and design rounds that differ by track.
- InstacartModerate coding, deep questions on classical ML and your own models, and grocery-flavoured ML design and experiment cases.
- WayfairA timed online assessment, then several business-flavoured ML case studies, one coding round and a behavioral round.
- eBayPractical Python and SQL coding, ML fundamentals and marketplace ML cases, with a heavy focus on your past projects.
- 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.