About

I build tools for Bitcoin forensics and responsible AI.

Through TrailBit Labs, I research the heuristics that blockchain analysts rely on to trace transactions — common-input-ownership, change address detection, timing analysis. The question driving this work: when we claim to know who sent bitcoin to whom, how confident should we actually be?

Through PLA Health, I'm building an iOS app that gives people health recommendations based on their biomarker data. Every recommendation carries an evidence grade — from L1 (backed by meta-analyses) down to L6 (preclinical only) — because when the stakes are personal, people deserve to know how much to trust what they're being told.

I'm increasingly interested in how the challenges of blockchain transparency mirror those of AI interpretability. Both fields are attempting to reverse-engineer meaning from systems that weren't designed to be understood. Both rely on heuristics that work until they don't. And in both cases, the gap between “usually correct” and “always correct” is where the real problems live.

I write about these topics in Bitcoin Heuristics Field Notes, a biweekly newsletter on blockchain forensics methodology — how heuristics work, where they break, and why it matters for compliance, privacy, and investigation.

By day I'm a Director at Deriv, where I lead the Cyprus office and work on AI product.

The problems I keep coming back to, regardless of domain: what assumptions am I making, how would I know if they were wrong, and what happens to the people using my tools if they are?

If you're working on Bitcoin forensics, AI interpretability, or responsible AI and want to compare notes — get in touch.