I have a couple of favourite ones but one I always think of immediately is a time when I was tasked with fixing some dashboard.

Sometimes students scores would be appearing as zero despite their average score being not zero.

Turns out it was some very simple averaging that didn’t filter out tests with a score of zero. That zero would then basically poison the entire calculation and only produce zero.

I don’t know how interesting that one is though.

I'd follow-up with questions on how you debugged that, how you'd do it now, what tooling you were missing, the testing strategy that led to this... I'd say it can be pretty interesting

The bug is never the interesting part. The follow on questions are where the data is. How did you find it? How did you fix it? What made it memorable? Did it change the way you code?