I’ve caught myself saying, “I think I like working on cars as much as I like racing them.” Reflecting on that, I’ve started to draw a parallel between my hobby and the systems engineering work I do in my day job.

I first tracked my BMW M235i at Harris Hill Raceway in San Marcos, Texas, learning the car’s limits on a lower-speed, technical track with long, sweeping corners. I was quick–I couldn’t quite reel in the Miatas and S2000s mid-corner, but I’d keep pace and reel them back in on the straights.

Then my “home” tracks became Laguna Seca and Sonoma Raceway—think higher cornering speeds, heavier braking zones, and significantly higher G-forces. The first system failure wasn’t mechanical, but a design limitation: oil pressure. After navigating the long, high-G carousel at Sonoma, I heard horror stories in the paddock of N55 engines like mine grenading because oil pooled on the far side of the stock pan during sustained loads, starving the engine.

An MMX Baffled Sump Plate solved that bottleneck, which immediately shifted the stress to the tires. I was destroying Michelin Pilot Super Sports every six months; no matter how I adjusted pressures, I’d wear the shoulders down to the cords. It was a very expensive feedback loop.

So, I stepped up to Yokohama Advan Neova AD09s and K-MAC Stage 2 Camber Plates. Dialing in -2 degrees of camber and utilizing stiffer sidewalls kept the tires from rolling over, both during warm-up and at peak operating temperature.

My lap times dropped dramatically, right up until the moment I felt my brakes fade for the first time. I cut my session short and returned to the pits to find discolored heat rings on my rotors and scorched pads. To close that loop, I swapped in Ferodo DS2500 pads and Zimmerman rotors, allowing the car to maintain its performance for a full 20-minute session.

The engineering side of racing is not too different from the systems engineering I do at work. You push a system to its limit, understand how its component parts are coupled, identify the new bottleneck, and then iterate on the design to move the goalposts.