I feel like the SQL round was harder than what I've seen at the big tech companies. Both were startups — one is a Series B company doing lending, the other is a Series D company doing SMB SaaS. The interview focus was similar at both: heavy emphasis on window functions, and rank vs dense_rank vs row_number came up without fail at both.
Both companies let you use window functions but wouldn't let you use joins — for things like rolling sums, lag(), lead(). Even if you got the right answer using a join, it didn't count.
The interviewers were disrespectful. At one company, halfway through the interview, the interviewer had the camera on and picked up a personal phone call. At the other, toward the end the interviewer just went off and did something else, didn't listen to me explain my coding logic at all, and had a flat-out sour expression the whole time.
Neither company cared much about edge cases or data quality handling. It felt less like they were evaluating from multiple angles whether you could actually do the job and solve problems, and more like they were just checking whether you could produce the full solution within the time limit, in the exact format they wanted. The two companies' businesses have nothing to do with each other, but their interview styles were basically identical.
I'm writing this for anyone prepping for startup interviews: being fluent from grinding problems is necessary, but just getting the answer isn't enough anymore — you also need to be able to flexibly use all kinds of window functions, not just left join. And at startups, not every interviewer has been trained to run an interview in a reasonably professional way.
Discussion
Loading comments…