Role track
Data Scientist
SQL for features, stats without vibes, and a holdout you would defend.
138 questions · SQL · Python / stats · Modeling · Experiments · Behavioral
Suggested sequence
1. SQL for features
Windows and point-in-time joins. Leakage is a join bug.
Scientist coding →2. The flowchart
Stats, experiments, and the holdout you will defend.
Scientist roadmap →3. AI / ML questions
Eval, leakage, and RAG as a pipeline.
AI questions →4. Behavioral loop
A time the experiment was invalid. What you stopped shipping.
Behavioral →
Stacks in this loop
Sample questions
hard · SQL · Stripe
Deduplicate to the latest customer row. Why not filter the window in WHERE?
Practice →medium · SQL · Capital One
A join exploded revenue. How do you find the grain mismatch in five minutes?
Practice →medium · SQL · Meta
Running total of daily revenue. Why is SUM(SUM(amount)) OVER (...) valid?
Practice →hard · SQL · Google
NULL = NULL is unknown. How does that bite a MERGE or an anti-join?
Practice →hard · SQL · LinkedIn
When do you use a covering index / clustering key versus rewriting the query?
Practice →medium · SQL · Snowflake
QUALIFY vs a CTE for window filters — when do you use each?
Practice →