From neuroscience labs to energy‑market risk models to industrial vision. The constant has been statistical thinking and stubborn attention to messy real‑world data.
Generalist · Sweden
Dominic Portain
Applied statistics, signal processing, and code that runs on small hardware.
I'm a cognitive‑science‑trained engineer who spends most of his time in the seam between models and matter, where a residual hints at what the equations missed and a sensor refuses to behave the way the spec sheet promised. I tend to leave the systems I build legible enough that someone else can pick them up later.
❦
Three ways in
This site is a small map of how I think, what I build, and what I'm slowly trying to put into words.
Around twenty detailed write‑ups and a longer archive: hardware tinkering, AI experiments, investigations into things I wanted to understand, and the occasional creative project.
Slow pieces about attention, design, and the texture of holding things. Less code, more language. There aren't many here yet; I write at the speed of contemplation.
The kinds of problems I’m drawn to
Things you can’t copy‑paste a solution for.
Models that misbehave
Real‑world time series, industrial signals, biomedical data: places where the residuals still tell a story. Model diagnostics, careful refits, and “why is this doing that?” investigations.
Code that touches devices
Embedded vision, smart cameras, pick‑and‑place machines, e‑paper panels, firmware to be reverse‑engineered. The kind of work where you can hear the system succeed or fail.
Making complexity legible
Turning models, pipelines, and hard‑won analyses into something colleagues and customers can actually use, trust, and maintain after I stop touching it.
Working on something that fits?
I'm open to roles and collaborations that mix statistics, AI, and hardware, and where the work has a clear line of sight to real users or real physical systems.