Let me start by saying you should try to avoid this interviewer if you can. He had a really heavy accent and I couldn't understand him at all, and when I asked questions partway through the interview, he just kept going silent.
The question was the IP address problem. I got to the third part, the CIDR block part, but when I ran it the answer was wrong. After the interview ended I thought about it more and realized the example he gave should iterate starting from the input address, while I was iterating from the start of the range.
OpenAI's coding environment is pretty tricky. First, when you run your code you're supposed to select "Run Test Case," not "Run Main" — but "Run Main" is the default. My interviewer wanted me to write my own unit tests, but what's really messed up is that the terminal output only shows a new test result after you click "Clear." It took me about five to ten minutes to figure out how running tests was even supposed to work, and my interviewer didn't help me at all — I figured out on my own that you had to click clear to see new output.
Haven't done the system design round yet — I'll post an update on that once I have.
Update: The system design question was to design a chess game. I hadn't prepared carefully for this one, just skimmed it a bit, since I figured it was more of a fullstack-style question and backend shouldn't get asked something like that. Turns out I did get asked it. The interviewer was decent — he asked things like how to track a user's remaining time, and how to handle recovery after a client fails, that kind of thing. I just kind of rambled through answers for all of it. I don't know how it turned out in the end. I guess OpenAI just has so many possible coding and system design questions that it's hard to prepare for all of them — you just have to leave it to luck.
Discussion
Loading comments…