On building at three boundaries simultaneously

Most people assume that working across power electronics, markets, and software means shallow depth in each. The opposite is true. The disciplines reinforce each other in ways that only become visible from inside all three.

The leverage point

A power electronics engineer who understands market microstructure has a different relationship to risk. Thermal resistance networks and signal decay have the same mathematics. A system that manages kilowatts of power and a system that manages a portfolio of signals are both control problems. The abstractions transfer.

The software layer is what lets you move between them at speed. Without it, you are slow. With it, you can run a simulation, analyse the output, and feed it back into a hardware design in the same afternoon.

What this actually looks like

It looks like reading Siegman's Lasers and Bouchaud's Trades, Quotes & Prices in the same week. It looks like writing Python to model real estate cashflows and then going to the bench to diagnose a switch-mode rectifier. The transitions are jarring at first. Then they stop being jarring. Then you notice that every domain has the same bones.

The discipline this requires

You cannot be romantic about any single field. The moment you decide that one discipline is your discipline and the others are supporting cast, you lose the leverage. All three have to be treated as equals.

This is the only interesting way to work.