This week was a wild ride at work. Solved an OOM disaster where 1.8MB of gzip'd json would decompress into over 600MB of JSON, which balloons into more when deserialized. And brought processing time down on some background jobs from 5 minutes to 3 seconds.
Would you believe me if I said linked lists are not sets and n^n algorithms are bad?
Incidents like these remind me why some tech companies interview for algorithm efficiency.
