Background & approach
From academic research and energy‑market analysis to embedded computer vision, the common thread is statistical thinking and getting models to behave on messy data.
Dominic Portain
Cognitive‑science‑trained engineer working at the intersection of applied statistics, AI, and hardware.
I work on systems where messy real‑world data, sensing, and physical devices meet — and try to make the modelling, code, and interfaces reliable enough that people can actually use them.
Current focus
Applied stats · Vision · Embedded
Real‑world data, close to hardware
I’m most useful where you have unusual sensor data, under‑documented pipelines, or models that don’t yet behave well outside a clean notebook.
See selected work →Where to go from here
This site is a small map of what I do, how I work, and a few things I’ve built.
Selected projects
A mix of applied statistics, signal and image processing, and hardware‑adjacent work. Not an exhaustive CV — just representative examples across domains.
Work together
I’m interested in long‑term roles and collaborations with real users, physical constraints, and enough autonomy to do deep, careful work.
What I tend to work on
Problems where you can’t just copy‑paste a solution.
Data & models
Messy time series and signals
Real‑world time series, industrial signals, and imaging data where the residuals still tell a story: model diagnostics, refits, and “why is this doing that?” investigations.
Hardware‑adjacent
Code that touches devices
Embedded and industrial vision, scanners and smart cameras, and workflows that connect software to pick‑and‑place machines or production lines.
People & interfaces
Making complexity usable
Turning complex analysis and systems into something colleagues, successors, and customers can understand and operate without a tour guide.
Good fit for your problem?
If you need someone to apply strong statistical and signal skills to real‑world data, close to hardware and actual users, I’d be happy to talk.