Nic Piraino

How I work

A schematic predicts how a circuit should behave, a simulation models it, and a datasheet defines what a part is supposed to do. The interesting part starts once the hardware exists and has to demonstrate what it actually does.

That has made me sceptical of designing by nominal value alone. Resistors have temperature and voltage coefficients. Amplifiers have offset, drift and bias current. Copper has resistance, connectors have drop, ground planes carry real current, and temperature moves nearly everything. I find those non-idealities more interesting than pretending they are not there.

The distinction I care most about is between calculated, simulated and measured. They are three different claims, and treating them as interchangeable is how a design ends up sounding better than it is. Where something on this site is calculated, it says so.

Experience

I worked as a Hardware Engineering Intern at Measurements International Ltd., a precision electrical metrology company, across summer terms in 2024, 2025 and 2026. My first term there was largely documentation and configuration control: modernising older schematics, producing revised board versions to match existing hardware, and generating BOMs, change logs and assembly instructions. That work taught me how quickly an electrical design becomes unmaintainable when nobody can tell which drawing matches the physical board.

My later terms moved into precision hardware itself: analog circuitry, current sources, impedance-related systems, PCB layout, component selection and characterisation. Working where measurement uncertainty is the product rather than an afterthought changes the scale at which imperfections matter. A resistor tempco that is excellent in a general-purpose circuit can dominate the error budget of a precision divider. A precision component does not make a precision instrument; the whole signal chain has to preserve it.

Selected work

A 10 V buried-Zener voltage reference, trimmed with a fixed metal-foil resistor network instead of a potentiometer, and characterised on an HP 3458A. A three-phase motor wound from scratch, where the winding sequence had to be redesigned after the first arrangement produced torque cancellation. A waste classifier that had to work on real rubbish rather than a curated demo set. On the software side, Tracesight and the other applications are where most of my testing discipline comes from.

Teams

I work on power architecture for Okanagan Rover Craft, UBCO’s entry to the Canadian International Rover Challenge. It is a 24 V LiFePO₄ system running battery → main fuse → contactor → high-current connector → individually fused branches, and it sits at the opposite end of the scale from a voltage reference: instead of microvolts, the questions are fault current, conductor sizing and whether every connection in the path can carry the load safely.

Before that I was on the electrical team for UBCO’s Aerial Robotics and Rocketry Club as head researcher for a helicopter FPV system, covering analog and digital video, transmitters, antennas and cameras. I also presented the project plan and budget to the CEO of Sanmina and secured platinum sponsorship for it, a reminder that a technically sound project still needs someone able to explain why it is worth funding.

Software and AI

I write a fair amount of software, but I treat it as another instrument rather than a separate career. A script that processes characterisation data or an application that automates a measurement workflow can be as useful as a physical tool.

I use language models heavily for development and research, and I have spent real time comparing models on coding, reasoning, context handling and cost rather than taking provider claims at face value. What I think actually matters is the workflow around them: specifications, bounded milestones, test suites, regression checks and explicit approval gates for anything irreversible. AI is very good at producing confident errors, so the value is not in generating code quickly. It is in being able to tell whether what came out is right.