Sharing my Rippling phone screen.
Phone screen, first round (coding):
The problem was something like a task scheduler, but the requirements were very complicated and there were a huge number of conditions. It's not hard, but the details explode. Roughly:
- Tasks have dependencies, and when outputting I had to output the dependency first and then the task itself.
- Tasks have various attribute requirements, such as importance/priority.
- I also had to filter out some of the data according to filter requirements.
This round was a bit unusual: the interviewer said I could use AI or not, and if I used AI I had to share my screen. I chose not to use AI at first and tried to understand the requirements and hand-write it myself, but there were too many scattered details, a human just can't catch them all, and I was too slow. Later I switched to AI, but there wasn't much time left, so I basically just had the AI generate everything and never really showed how I work together with AI.
Result: rejected. Looking back, for a problem with this many details, you should either use AI from the very beginning and clearly show the collaboration process, or, if you hand-write it, you have to catch all the requirements extremely fast. Switching halfway is the worst of both.
Discussion
Loading comments…