Engineering before product management.
The industries changed: plasma physics research, electrical equipment engineering, telecommunications infrastructure, software automation, AI products. The underlying habit did not. Understand the system, find the bottleneck, build a better solution.
What was not working
Before product management, the work was hands on: physics experiments, electrical hardware, and field infrastructure, each with its own constraints, materials, and failure modes.
What the user actually needed
Every environment rewarded the same habit: understand how the system actually behaves, find where it breaks down or wastes effort, and build a concrete fix. A curve fit, a tolerance adjustment, an automation script.
What I owned
- NYU Tandon Plasma Physics Lab: designed Paschen Curve experiments isolating the effect of magnetic field orientation on gas discharge breakdown voltage, and built the lab itself from an empty facility to operational readiness.
- Decom Electrical: co engineered a 600mm wide SF6 gas insulated switchgear system, producing 3D models and assembly drawings in SOLIDWORKS and AutoCAD from concept through prototype.
- A.T. Kearney: analyzed customer and store performance data for a major retail client, contributing to a 15% sales increase.
What had to be true
- Physical hardware has no ship a patch later. Tolerance and assembly decisions had to be right before prototype build.
- A brand new lab meant building infrastructure (vacuum systems, safety integration) before any experiment could run at all.
From physics to hardware to infrastructure
Each role sharpened the same instinct in a different medium: MATLAB models for curve fitting in the lab; DFM/DFA and tolerance review for switchgear assembly; and later, workflow analysis for telecom infrastructure and AI systems.
What shipped
- A working plasma physics lab, built from equipment acquisition through vacuum system setup and safety integration.
- A 600mm wide SF6 gas insulated switchgear system carried from concept to prototype assembly.
How it was built
- Not applicable in the software sense. The throughline is methodology: instrument the system, isolate the variable, fix the bottleneck.
Trade offs I made on purpose
- Carried a bias toward hands on validation, build the experiment, build the prototype, test it, into later product and engineering work.
Takeaways
- Systems thinking transfers across mediums. Plasma physics, electrical hardware, fiber networks, and AI agents are different systems with the same underlying discipline.