R&D work on automatic transmissions and early hybrid propulsion systems.
Engineering and simulation work on automatic transmissions for medium- and heavy-duty commercial vehicles. Core work involved powertrain control, mechanical systems integration, and reliability engineering for demanding real-world duty cycles.
Later projects centered on flywheel hybrid transmissions for commercial utility vehicles — exploring regenerative energy storage, power management, and hybrid efficiency in applications where recapturing braking energy offered strong real-world gains (delivery, utility, vocational).
The starting point was a flywheel hybrid simulation developed by engineers at the GM Technical Center for consumer automobiles. The simulation was written in PL/1 — a mainframe-era language still in active use at GM for complex engineering models. The work involved adapting that model to accurately simulate heavy commercial vehicles: adjusting for different mass profiles, duty cycles, regenerative energy storage behavior, and power management characteristics.
This kind of work — taking someone else's complex simulation, understanding it deeply enough to safely extend it, and verifying that the modified model produces trustworthy results — is the same discipline that runs through everything built since. The tools change; the problem of inspectable, modifiable systems doesn't.
Collaboration with GM Tech Center engineers on the flywheel hybrid simulation created visibility into broader advanced technology initiatives at GM. Through those connections, learned of a new Expert Systems project being undertaken with Teknowledge — a pioneering knowledge engineering firm in Silicon Valley backed in part by GM investment.
Applied for the position, was selected, and relocated to California to join the Teknowledge team. That move marked the transition from mechanical and powertrain R&D into full-time knowledge engineering and symbolic AI work during the peak of the 1980s expert systems boom — and the true launch of a 40+ year software career.