Skip to content
Jarryd Aubert.

I build to understand what I test.

To test software well, I wanted to feel what building it is actually like. So I've shipped my own apps and sites, start to finish, and held them to the same scrutiny I'd give anyone else's.

See the projectsWhy I build

01 Projects

What I built, and tested properly.

Live

PayeTax

A live UK PAYE, salary, and dividend calculator.

What I built
I chose it because it gave me something a tester wants: a clear right or wrong answer. The work forced me through messy inputs, including Scottish rates, student-loan thresholds on total income, and salary/dividend choices.
It matches my own payslip to the penny.
What it demonstrates
Correctness under real-world inputs, not happy-path ones.

Visit PayeTax

How I tested it, and where that testing stops →

Live

ProsePal

A live message-writing app.

What I built
I built and shipped it as a real app, then used it to explore the harder problem of judging output where there is no single right answer.
A message can be grammatically perfect and still completely wrong for the moment. No assertion catches that, which is exactly the problem I am working through.
What it explores
How do you test something when correctness is a matter of judgement, not a value you can assert against?

Visit ProsePal

02 The Why

I build to understand what I test.

I wanted to feel firsthand what building software is actually like, from buying a domain and setting up a repo to deploying, publishing, watching analytics, and living with my own bugs.

Going through that pain changed how I test, because now I understand more of where software actually hurts. The awkward handoffs, the ambiguous decisions, the hidden edge cases, and the small release details all feel different when my own name is on the thing.

Writing

Notes on testing, building, and shipping software that holds up.

I tried to prove my tax calculator was correct. I mostly proved it was consistent.

AI demos reward speed. Shipping rewards judgement.