Three-quarters of a mile. Every morning. Or a shuttle that showed up whenever it felt like it. For two to three years, I counted down the months like a kid counting down to Christmas, except the prize was a parking spot in the underground garage, not presents.
The estimate that got a project reassigned
Years later, I got handed the exact project that had once controlled my own morning walk. CalPERS needed to modernize its employee parking and waitlist system, centralizing vehicle records, garage assignments, and real-time waitlist position for somewhere between 1,100 and 1,500 tracked employees, all of it running on fragmented spreadsheets and manual processes that were never going to hold up under a coming surge in office occupancy.
A team of four developers looked at that scope and told the division chief, three levels above my manager, that it would take six months.
He didn’t push back on the number. He didn’t ask them to move faster or cut corners. He pulled the entire project away from all four of them and handed it to one person.
Why one person instead of four
This wasn’t a snap decision made out of frustration. It came from an existing track record: a reputation, already established well before this project, for being the person who actually finishes things rather than the person who reports steady, reasonable-sounding progress on something that never quite lands. Four developers, given six months, tend to produce meetings, sprint estimates, and status updates. One person who’s actually going to finish it produces the finished thing.
With a return-to-office mandate on a fixed timeline and an operations team that needed a working system now, not eventually, that distinction stopped being a matter of personal style and became the only thing that actually mattered to leadership.
What got built, and how fast
Three calendar weeks. Not six months. A full system: employee and vehicle tracking, support for roughly 700 active garage vehicles, variable-length carpool passenger lists, and a waitlist ranking system built directly into a SQL query rather than calculated after the fact in application code, so a real-time position could be computed efficiently without unnecessary processing overhead. Admin tooling for managing users, parking lots, and vehicle types. A lightweight confirmation system on registration, deliberately minimal, since parking operations wanted to retain manual control over final prioritization rather than hand that judgment entirely to automation.
It went live in May 2021, ahead of the return-to-office deadline it was built for, and then it just kept running. No major refactor. No rescue project years later. It scaled cleanly through the actual wave of return-to-office changes that followed, changes it hadn’t been specifically designed to anticipate, because the underlying design was clean enough to absorb them anyway.
Where the empathy actually mattered, and where it didn’t
I’d lived on the losing end of this exact system, waiting years for a spot with no real visibility into where I actually stood in line. That firsthand frustration shaped a real decision in the final design: people could see their actual position, not just wait and wonder, because I remembered wanting that myself and never getting it.
But that’s not why the chief handed me the project. He didn’t know about my personal parking history when he made that call. He made it because he needed something to exist by a date that wasn’t moving, and he already knew, from prior work, exactly who was going to make that happen without three more status meetings about it first.
The actual lesson
“Experienced” and “fast” aren’t the same claim, and this project is the clearest example I’ve got of the difference. Four experienced developers produced a six-month estimate. One person produced a working system in three weeks. The gap wasn’t talent, and it wasn’t even really about parking. It was about which kind of person a leadership team reaches for when a deadline genuinely can’t move: the one who explains the plan, or the one who’s already finished it by the time the explanation would have been due.