@jonathan.interviews: This SQL problem showed up in my senior and staff data scientist interviews at some of the biggest tech companies in the world. Nearly identical questions with slight wording changes. The task: track monthly active customers by company, split by premium vs. non-premium, and compute month-over-month changes. It tests three things that separate strong candidates from average ones — conditional aggregation inside COUNT DISTINCT, partitioned window functions with LAG, and knowing when to use a CTE to stage your aggregation before applying those windows. If you can’t do all three cleanly in one query, that’s the gap to close. #sql #datascience #dataanalytics #datascientist #techinterview

Data Science Interviews

Open In TikTok:

Region: US

Sunday 19 April 2026 00:26:11 GMT