Nothing is wasted.
Late last year I was building an algorithm to detect hidden biases. The kind that sits beneath the surface and shapes everything downstream without announcing itself.
I hit a wall. But it wasn’t a technical one.
Who would have a use for this beyond academia? Who actually needs this enough to fund it?
I didn’t have a good answer. So I put it down.
I’m an Iraqi-British materials scientist who became an AI developer. I started as a creative. Non-linear is how I work.
When you put something down, you don’t put down what it taught you.
That bias algorithm gave me an intuition I didn’t know I needed. It made me ask: what gets recorded, when, and by whom? Whose environment gets to be the default? That thinking became foundational to the AI audit trail I’ve been building - professional infrastructure for AI assurance that has to work in the real world, not just the compliant one. It reshaped how I think about drift.
Alongside that, I’ve been developing a PWA for studying date palms and palm ash - Iraqi environmental heritage encoded as research software. Sounds like a pivot. It isn’t. It’s the same question in a different material.
While I was in Iraq, I was asking relatives about their relationship with technology. What I understood wasn’t in any compliance document: the gaps aren’t just in the data. They’re in the security. In what it costs someone if their information is exposed where institutional protection isn’t guaranteed.
The EU AI Act, for all its ambition around data governance and risk tiers, was written for environments with reliable infrastructure and enforceable law. Article 10 comes closest to what I was juggling with. But closest isn’t enough when you’re building outside the dominant frame.
So I slowed down. Not because I couldn’t ship. Because responsible software for the Iraqi environment had to meet security needs before deployment. Full stop.
There’s a market question I haven’t resolved: who funds public-interest AI assurance for underrepresented environments? Not academia alone. I think it lives at the intersection of AI governance, environmental research, and communities who need infrastructure built with them rather than for them. That’s where I’m working.
My father made this recently. Iraqi marquetry, mother of pearl inlaid in wood. Islamic geometry operates on a principle: everything comes from the centre. Materials science is my centre point. Data security and purpose aren’t distractions, they’re part of the inlay. You can’t complete the pattern without them.
The bias algorithm, the audit trail, the palm research, the conversations - a scattered portfolio from the outside. From the inside, a single question asked across different materials:
What does it mean to know something responsibly in an underrepresented context?
Nothing I’ve built has been wasted.
