00:11 · MayaHow did you measure the win — what was it before, and after?
00:18 · MarkThe list view rendered in about 240ms; after the change it was closer to 90ms, measured on the same dataset.
01:02 · MayaWhat was the actual bottleneck?
01:09 · MarkWe were re-rendering the whole list on every keystroke. I memoised the rows and moved filtering off the main thread.