Whatnot rejected me after the sandboxing question stumped me

Whatnot rejected me after the sandboxing question stumped me

Had my phone screen with Whatnot last week, they packed coding and system design into the same hour long session.

First problem was Count Singler Role Co-occurrences, a hash table grouping problem where you track which roles a singer performs and count co-occurrences within a set number of years. Took a minute to get the sort order right on the output but landed a working solution.

Then there was a quick follow up on handling duplicate roles for the same singer, mostly just talked through the approach since time was tight.

SD half was Design LeetCode, basically problem submission, code execution, and a leaderboard. Interviewer really wanted to hear about running arbitrary user code safely, sandboxing, containerized runners, that whole area. I hadn't thought that deep about it going in and it showed.

Rejection came two days later, no specifics. Bar felt high honestly, feels like they want clean fast code and full follow up coverage, a working answer alone doesn't cut it.

Anyone else find the SD half tougher than the coding half here?