About
Hi, I'm Natalie.
I'm a London-based fraud and risk professional with a background in investigations, financial judgement and complex operations. I also build independently using AI-assisted development.
I spent three and a half years at Octopus Energy, starting in billing, refunds and complex operations before moving into Revenue Protection investigations. The through-line was always the same: work out what is really going on when the data is messy or doesn't add up, weigh the risk, and make a decision I can defend.
The building side
Outside my main career, I build things. What started as using AI for research developed into building real products and tools. I began writing specifications, defining rules and edge cases, directing implementation, testing behaviour and iterating until the result actually worked.
Since then I've shipped an iOS consumer app and built trust-and-safety tooling, automated trading systems, an operational dashboard and a Spotify library auditor.
How I use AI
I use AI-assisted development extensively, but I don't start with “build me an app.” I start with the problem: what the system should do, what it should never do, what it can trust, and what happens when data is missing. From there I write the specification, direct the implementation, review what comes back and test it until I trust the result.
I'm still learning the technical stack as I build, and I'm deliberate about not claiming fluency I haven't earned yet.
How I think
Curiosity is usually where it starts. A messy music library becomes a question about detecting duplicates across re-releases; a subscription app becomes a question about spotting abuse without punishing users who are simply unusual; a pile of scattered tabs becomes a question about what one clear operational view would look like.
I like following those questions — and I find it just as useful when they end up proving the original idea wrong.
What comes next
I'm looking for work where curiosity, judgement and problem-solving genuinely matter — in fraud, risk, investigations, trust and safety, complex operations or somewhere adjacent. I'm especially drawn to teams comfortable questioning how things currently work, and curious about how technology and AI can make good decisions better without replacing the judgement behind them.