Snowflake · Primly Community

Snowflake coding interview / online assessment: format and difficulty, honest take

infra_ines · 4 replies

Went through the Snowflake online assessment and technical phone screen last month for a senior SWE role. A few people asked me to write this up so here it is.

Online assessment

Two problems, 90-minute window. You get both at the start and can switch between them. Platform is HackerRank, not Coderpad. Auto-grader runs test cases.

Difficulty: I'd call it one medium and one medium-hard. The medium was a dynamic programming problem that wasn't super tricky if you spotted the pattern. The medium-hard involved parsing a structured data format with some edge cases that were easy to miss.

Tips: Read both problems before starting. Figure out which one is faster to solve first. The test cases aren't exhaustive. Write your own edge cases before submitting. They're looking at code quality in some OAs, not just pass/fail. Write clean code even if it passes.

Technical phone screen (for context)

This was separate from the OA in my loop. One problem, 60 minutes, Coderpad. Interviewer is live. Problem was graph traversal, nothing exotic. The last 15 minutes were an explicit discussion: time complexity, space complexity, how does this break at 100x scale.

The live round felt more like an actual conversation. I prefer it honestly because you can explain your reasoning. The OA is pure output.

Overall difficulty vs. other companies

Harder than a lot of late-stage companies I've seen. Roughly on par with Databricks or Confluent in terms of coding bar. Not at the FAANG top end (Google/Meta hard rounds) but not a free pass either. If you're LC grinding: comfortable with mediums and exposure to graphs, trees, and DP will cover most scenarios.

One more thing: they move fast. Got OA within 2 days of recruiter screen, phone screen scheduled within a week after that. If you're not ready to interview immediately, say so on the recruiter call.

4 replies

alex_design

do they actually look at code style in the OA or is that just what people say to feel better about writing comments

brand_ben

honestly don't know for certain. i write clean code by habit so I can't isolate that variable. but I've seen companies say they do review it for senior roles. not worth risking it either way.

ux_uma

for the data science track the OA I did was different: one SQL problem and one Python data manipulation problem, no graph stuff. so the coding format may vary by role/team.

firsttime_mgr

the 'read both problems first' tip is SO important. I've wasted 30 minutes on a harder problem before realizing the other one was trivial.