Turns out writing about yourself is harder than shipping a distributed system. Not going to pretend this is a pitch — it's closer to a diary entry about five years of finding out, over and over, why things break under real load.
My first year, every bug felt personal. Nobody tells you in college that the real skill isn't syntax — it's caring enough to trace a request through an entire system until you actually understand why it broke.
Seven people, one product, and nobody else to ask when the backend broke at 2am. I built the whole API layer myself, and learned the difference between code that works and code you can walk away from and trust.
Moving a lending platform's database while it stayed live — real money, real transactions, still running — taught me what "careful" actually means. Four hundred thousand records, and I didn't relax until the very last one landed.
Some bugs only exist for a few milliseconds a day, under exactly the wrong conditions. I spent months hunting race conditions that only showed up under real traffic, and came out the other side liking concurrency instead of fearing it.
These days I keep a system serving more than a dozen tenants from tripping over itself — and I've started teaching Claude the parts of the job I already know by heart, so I can spend more time on the parts I don't.