how i got here
There was a stretch of years when this work had gone dull on me. A live system the whole school depends on, demand that always outran what two programmers could deliver, and underneath it a layer of legacy security debt nobody had the spare capacity to touch. The queue never got shorter. Not because the problems were beneath me — because there was never enough throughput to get ahead of them, and work you can never get ahead of stops feeling like building anything.
AI changed the arithmetic. Not by doing the thinking — by multiplying what one careful engineer can hold, produce, and verify in a day. For the first time in years, output overtook demand. And the surplus is the interesting part: once the ticket queue stopped consuming everything, there was finally room to fix how we work, not just what's broken. I used that room to bring version control to the team — a Git workflow and SOP now run by our two programmers, with a release step our HOD and manager can actually see and approve, instead of deployments nobody could trace.
Working with AI on deep system fundamentals, one engineer now delivers what used to take a team — at a quality I'd defend line by line.
That's the part I didn't expect. The AI alone doesn't do this — it works because it's paired with someone who knows where the data lives, what breaks, and why every odd rule in a twenty-year-old system exists. Fundamentals are what make the leverage safe to use. And after years of feeling buried by this system, I genuinely enjoy the work again.