VMware · Primly Community

VMware data scientist interview: SQL, stats, and case questions, what the loop actually looks like

ds_dmitri · 5 replies

Wrapped the VMware data scientist loop in January 2026 for a role on their customer success analytics team. This was my first enterprise-infra DS interview and it had a different texture than what I'd seen at consumer or fintech companies. Sharing the breakdown.

SQL round: Definitely present. One dedicated 45-minute SQL round. HackerRank or a shared doc, I did mine in a doc. Questions were intermediate to advanced. Window functions came up immediately: running totals, ranking within groups, lag/lead for month-over-month comparisons. One multi-table join with aggregation. Not just "select from where group by" stuff.

They also asked me to interpret a slow-running query and identify the bottleneck. That was a bit unexpected. Have at least a passing familiarity with query execution plans and indexes.

Statistics / methods round: About 30 minutes of this within a longer round. Topics: A/B test design, how to determine sample size, how to handle imbalanced groups, how to present results to a non-technical VP. One question about cohort analysis for customer retention. They did not ask me to code a model live. It was more conceptual.

Case / product sense: "A key customer segment's support ticket volume has risen 40% over the last two quarters. Walk me through how you'd analyze this." This felt like a mix of DS and PM. They wanted data sources, hypotheses, an analysis plan, and a narrative for how you'd present findings to leadership.

What I didn't see: Machine learning depth (no model architecture questions, no feature engineering deep dives). This role was analytics-heavy, not ML engineering-heavy. Know your role type before you prep.

Comp data point: offer was for remote, mid-senior DS. Base around $155k, bonus target 10%, RSU grant in Broadcom stock over 4 years. Total first-year closer to $185k with vesting.

5 replies

analyst_ana

The slow query / execution plan question is one I always get tripped up on. Do you have resources you used to prep for that?

brand_ben

Honestly just EXPLAIN ANALYZE in Postgres and reading a few articles on index usage. It's not deep database engineering, they just want to see that you know indexes exist and that sequential scans are expensive on large tables.

ux_uma

$155k base for remote DS mid-senior in 2026 is on the lower end of market but not unreasonable for enterprise infra. The AVGO stock has been moving so the RSU component is the variable.

returner_ren

The A/B test / sample size question comes up in almost every DS interview I've heard about. Is it the full power calculation walkthrough or more of a gut-check question?

firsttime_mgr

More gut-check + reasoning. They didn't want me to derive the formula. They wanted to know: what factors go into deciding sample size, what do you do if you can't hit the required N, and how do you think about running time vs. confidence trade-offs. Conceptual, not computational.