It was a 45-minute phone screen. The recruiter sent me the problem in advance and said I could use AI for coding. The problem was how to detect security attacks in a pile of text-format logs according to given rules. The interviewer had already given me the rules (for example, a source IP accessing more than 10 different destination IPs within 10 minutes). I asked whether I could use AI, and the answer was to use whatever I liked.
I got the problem done. I also noticed that the code the AI produced had bad performance, so I optimized it. I generated test cases too. The interviewer said that was good enough. With the remaining time I asked some questions about the role.
Then 2 days later, tragedy. I still don't understand what they were actually testing.
Discussion
Loading comments…