interviewing at NYT for a data scientist role in 3 weeks (reader analytics team). would love to hear from anyone who's been through the loop recently, especially on the DS/analytics side. a few specific things i'm trying to figure out: how SQL-heavy is the technical round? like, is it clean analytical SQL or do they get into window functions, complex joins, performance stuff? do they give a take-home or is everything live? how much product sense do they test vs. straight modeling?
drop whatever you know. i'll report back after.
4 replies
de_derek
went through a data eng round there about 6 months ago, different role but adjacent. SQL was definitely in there and they did ask some window function stuff: rank, lag, finding gaps in time-series data. not impossible but they wanted to see clean query thinking, not just whether you know the syntax. there was also a case-style question about how you'd measure the impact of a paywall change on reader behavior.
analyst_ana
i applied to a similar role last year and got a take-home. it was a dataset of article engagement metrics and they asked you to identify trends, build a simple model if applicable, and write up findings as if presenting to a non-technical editorial lead. that last part is the key: they want communication, not just accuracy.
ds_dmitri
this is super helpful, the non-technical write-up angle is something i wouldn't have prepped for. thank you.
ux_uma
from a research side (i know, different function): they care a lot about how you frame impact in terms of reader and subscriber outcomes, not just model metrics. even for a DS role i'd bet framing around "what does this mean for retention or engagement" goes over better than purely technical framing.