I no longer remember the technical screen. I felt reasonably good afterward, but I did not hear anything and assumed I had failed, until three weeks later they invited me to the onsite.
There were four rounds in total.
The AI round was the one discussed in other posts: use my local IDE and AI. The requirements appeared on HackerRank, and I had to build a local server from scratch. It was roughly about implementing a refund-request feature. It only needed to run locally, with no frontend—pure backend. I felt I wasted some time at the start, but I did get it running in the end. This round felt barely okay.
The engineering-manager round: Not much to say, just behavioral and situational questions. None was especially unusual. It felt fairly average.
The system-design round started with about half an hour of a deep dive into a previous project, followed by an hour designing a donation system. I felt reasonably good about this one. (One question: why do some people get the half-hour project deep dive and others not? What determines that?)
The debugging round was the random-Dasher question many people have posted about. I initially felt this was where I failed. The bug looked simple, but I only got it working at almost the last moment, and I did not feel my fix was very good. There was no consistent-hashing follow-up, but I was asked what to consider when putting it into production.
Overall, it seems every round was only barely okay. I probably failed in the engineering-manager round. The interviewer was very nice. But when she asked about a production bug, I could not think of a good example and gave a general answer. I initially thought I failed in debugging, but it was probably the manager round.
Discussion
Loading comments…