Great write-up. The biggest lesson I’ve learned over the years is that humans are notoriously terrible at guessing where bottlenecks will happen. I can't tell you how many times I spent hours optimizing a nested loop, only to profile the application later and realize 95% of the latency was just a poorly indexed database query or a slow third-party API. If you aren't measuring with actual profiling tools, you aren't optimizing—you're just guessing.