Three candidates for Wayfair's machine learning scientist track describe an online assessment, ML case studies, coding, and behavioral questions. Prepare for churn and recommendation cases, NumPy and SQL coding, and project-ownership stories.
$155K
Avg. Base Comp
$220K
Avg. Total Comp
Not reported
Typical Rounds
Not reported
Process Length
Wayfair's AI Research Scientist interview, as three candidates describe it for machine learning scientist roles, mixes applied ML case studies with hands-on coding. Treat the evidence as a small sample, because the three reports differ in level and format.
Two candidates began with a HackerRank assessment. One described two medium LeetCode-style problems plus nine multiple-choice questions on statistics and basic data structures. The other described a mix of coding and multiple-choice questions with two weeks to finish. A third candidate began with an HR screen and a manager case study about device tracking.
ML case studies were the most common thread. One candidate reports a customer churn prediction discussion covering features and validation, followed by an onsite case that pushed harder on assumptions and trade-offs. Another recalls search-query classification, recommendations for users who visited 30 days earlier, and probabilistic device graphs among the case problems referenced. One candidate found the prompts deliberately vague and had an interviewer shut down clarifying questions, so practice structuring an answer from a thin prompt and naming the formulas you rely on.
Coding varied. One candidate had 60 minutes for two medium algorithm problems. Another was asked to implement a machine learning algorithm from scratch in NumPy and ran out of time. A third was asked a SQL question and also mentions a heap problem on stone weights and an interval scheduling problem among the problems referenced. Practice NumPy fluency, SQL, and priority queues.
Behavioral rounds asked about owning a project end to end, handling tough situations, and cross-team collaboration. Prepare concrete project stories with your own decisions and results. All three reported no offer or a declined offer, and one reported a roughly two-week span from assessment to final round.
Synthesized from 3 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 Wayfair process.
I went through Wayfair's AI Research Scientist interview process starting with a HackerRank online assessment. The OA had two parts: two medium-level LeetCode problems and nine multiple-choice questions covering statistics and basic DSA concepts. The recruiter sent some prep guides, though they varied in guidance, which was a bit confusing at first, but the core expectation was clear enough.
After passing the OA, I moved to a technical screen focused on an ML case study. The interviewer walked me through a customer churn prediction problem, asking me to think through the approach, potential features, and validation strategy. It felt more like a collaborative discussion than a grilling—they wanted to see how I'd structure the problem.
The virtual onsite had three components back-to-back. First was a deeper dive into another ML case study with a different problem statement but similar conceptual territory. The interviewer pushed harder on assumptions and trade-offs this time. Then came a behavioral round that honestly felt oddly framed—they asked me to describe handling an entire project end-to-end, which felt disconnected from my 1–2 years of experience at that point. The question assumed more seniority than the role seemed to target.
Finally, there was a coding round with 60 minutes to solve two medium-level LeetCode-style problems. The problems were straightforward algorithmic challenges, nothing domain-specific. The whole process from initial OA to final round took roughly two weeks, and while the structure was clear and the interviewers were professional, I got the sense there was some friction between what different parts of the team were evaluating for. I ended up declining the offer.
Prep tip from this candidate
The behavioral round asked for end-to-end project ownership on a role that expects 1–2 years of experience—that mismatch suggests you should come prepared with concrete examples of projects you've led or significantly owned, even if they're smaller in scope, and be ready to articulate them clearly. Focus also on drilling the ML case study frameworks (churn prediction, feature engineering, evaluation metrics) since they appeared twice with increasing depth.
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 Wayfair
How would you set up this test?
| Question | |
|---|---|
| Prime to N | |
| Network Experiment Design | |
| Hurdles In Data Projects | |
| Testing Price Increase | |
| Closest Key | |
| Addressing Data Quality Issues | |
| International e-Commerce Warehouse | |
| A/B Testing a Checkout Button Change | |
| Youtube Recommendations | |
| Client Solution Pushback | |
| Stakeholder Communication | |
| Summing Numeric Strings | |
| k-Means from Scratch | |
| Friends Over Engagement | |
| Your Strengths and Weaknesses | |
| Free Shipping Mention Test | |
| Why Do You Want to Work With Us | |
| Statistically Significant Test | |
| Marketing Dollar Efficiency | |
| Parking Application System Design | |
| Listings Recommendation | |
| 2nd Highest Salary | |
| Merge Sorted Lists | |
| Experiment Validity | |
| Scrambled Tickets | |
| P-value to a Layman | |
| Bagging vs Boosting | |
| Compute Deviation | |
| Nearest Common Ancestor |
Synthesized from candidate reports. Individual experiences may vary.
Two candidates report a HackerRank assessment mixing coding with multiple-choice questions. One described two medium problems and nine statistics and data structures questions. Another candidate reports an HR screen about background and interest in Wayfair, so the opening step may vary.
Candidates report a first case study with a manager or interviewer. Topics included customer churn prediction, device tracking, and an ML system design for a website feature. Expect to discuss approach, features, and validation, and the style may range from collaborative to vague.
Two candidates report further case studies in the onsite loop, with more pressure on assumptions and trade-offs. A third mentions search-query classification, recommendations, and probabilistic device graphs among the problems referenced. One candidate says an interviewer discouraged clarifying questions, so be ready to state your assumptions explicitly.
Coding formats differ by candidate. Reports include two medium algorithm problems in 60 minutes, implementing an ML algorithm from scratch in NumPy, and a SQL question. One candidate also mentions heap and interval-scheduling problems among those referenced. Preparation should cover several formats rather than one.
Candidates report behavioral questions about owning a project end to end, handling difficult work situations, and collaborating with other teams. One candidate found the ownership expectation high for a junior profile, so prepare concrete examples with scope, decisions, and results.
