Snowflake Software Engineer Interview Experience — Two Onsite Coding Rounds, Rejected Over a Tricky Input Assumption

Two onsite rounds, mid-level, both coding.

Both rounds were repeat questions I'd seen before, but they happened to be from the few I hadn't practiced. A bit unlucky. Overall the interview experience was fine — the interviewers were all nice, and they asked me a lot about how I use AI in my day-to-day developing cycle.

1. Wiki Hopper — basically a webpage has multiple links that let you jump to other pages. The question was how to return all the pages. I started with DFS, and after I finished writing it, the interviewer asked me what problems DFS would have, so I said it could be optimized to BFS. The core of the question was really about why BFS is better than DFS here — other writeups on the forum cover that part. The final follow-up was how to process BFS in parallel with multiple threads.

2. Snow Cal. This one was a bit tricky, because I initially assumed the input would be given one at a time, but once I started writing I realized that wasn't right — after I asked the interviewer, it turned out the input was actually a list. He probably expected me to ask about that proactively from the start; I guess after doing so much OOD I subconsciously assumed inputs come in one by one. With a small hint near the end I more or less got it written out. I didn't have time for the follow-up. I think I failed this round because I wasn't strong enough on it.

Ugh, I didn't get many interviews, and I prepared a lot of repeat questions that just never came up.