interviewing at DoorDash for a senior data engineer role in three weeks. found some old threads but nothing recent. would love to hear from anyone who's gone through their data loop in the last 6-12 months.
specifically curious about: is it still heavy on SQL and system design? do they give you a take-home or is it all live? any sense of what 'data platform' vs 'analytics engineering' roles look like differently in the loop? and is the behavioral component actually a separate round or woven into everything?
drop whatever you have, even partial data points are useful.
4 replies
ds_dmitri
did a DS loop there about 5 months ago. no take-home for DS, all live. two rounds: one was pure SQL (window functions, some tricky self-join, a metrics calculation scenario), the other was more ML-flavored, 'how would you approach building a model for driver ETA prediction.' they cared about feature engineering and how i'd validate the model more than specific algorithms. behavioral was a separate 45-minute round.
de_derek
super helpful. the SQL window function emphasis tracks. any sense of how hard the SQL was? like are we talking intermediate or 'you should have been a database in a past life'?
ds_dmitri
intermediate-to-hard honestly. one of the problems required nesting a window function inside another aggregation and getting the order of operations right. the kind of thing that trips you up if you're not writing complex SQL regularly. not impossible but not a warmup either.
analyst_ana
i interviewed for an analytics engineering role last fall. they gave me a live SQL problem AND a 'design this data model for merchant analytics' problem. the data model one was more open-ended than i expected, they wanted me to think through grain and slowly changing dimensions. for behavioral they really wanted examples of cross-functional collaboration and impact.