Ledger platforms. Tax compliance engines. Reconciliation pipelines. Cross-border payment infrastructure. Across blockchain and traditional fintech — at NASDAQ-listed scale.
18+ years in engineering. 10+ across payments, fintech, and crypto. I lead teams that own mission-critical financial infrastructure — the kind where a single miscalculated entry triggers a regulatory inquiry, not just a bug ticket.
At Coinbase, I owned the ledger modernisation roadmap — automated accounting across hundreds of fund flows, compressed settlement timelines by an order of magnitude, and rapidly scaled the tax engineering function to meet a regulatory deadline, delivering IRS digital asset reporting for 70M+ US users. I also pioneered AI-first engineering in a regulated environment — the majority of engineering work authored via Claude with compliance guardrails in production.
At Nium, I scaled an engineering org several-fold, delivered cross-border payment infrastructure from zero across 100+ currencies and hundreds of bank accounts, and built in-house reconciliation that replaced a six-figure vendor — all at enterprise-grade reliability, driving 85% of the company’s global payout revenue.
Before that, I founded Trice Technologies — an eLearning product that reached 100K+ licences and exited at 10x.
The systems I build are precise. The teams I build are not. People ship great work when they’re trusted, supported, and given a clear runway — not when they’re managed by the minute.
Every architectural choice has a business consequence. I’ve learned to think in revenue, regulatory risk, and time-to-value — not just system design.
The best engineering decisions are the ones a CFO would also approve of.
I hire for judgment, coach for growth, and create environments where people can do their best work — even on regulated, mission-critical systems.
The hardest problems get solved by teams that feel safe, not teams that feel watched.
Data contracts, audit trails, edge cases, dollar-impact dashboards — the details matter because the consequences are regulatory. But details are the team’s job to own.
My job is to set the bar, not to inspect every line of code.
Engineering Review Charters. OKR-driven planning. Shift-left testing. Self-serve platforms that eliminate engineering tickets entirely.
The frameworks I build meaningfully lifted Nium’s engineering throughput — without burning anyone out.
Multi-year technical roadmaps that survive reorgs, leadership changes, and scope shifts. The Coinbase ledger transformation and Nium’s reconciliation platform both took years — and held their shape because the planning was real.
Predictable delivery is the goal. Not heroic recovery.
Zero audit failures. Enterprise-grade reliability. T+1 settlement. None of these were achieved through crunch — they came from clear ownership, audit-ready-by-default design, and teams that trust the system.
Calm engineering isn’t boring. It’s how you protect billions.
Every metric below was delivered in a regulated environment, audited externally, or scaled to production. No vanity numbers.
Two financial infrastructure platforms. Two regulated environments. Both built from the ground up — one in crypto, one in cross-border payments.
Four case studies on building and leading engineering teams that own mission-critical financial infrastructure — from ledger transformations to cross-timezone org design.
Detailed case studies with context, decisions, and outcomes.
Deep domain expertise across financial infrastructure, compliance, and engineering — not a generalist learning payments on the job.
Straight answers on the systems, the domain, and leading engineering in the AI era.
It's the engineering behind systems where money moves and every number has to be provably correct — ledgers, reconciliation engines, settlement pipelines, and the integrations that feed accounting and tax. The bar is different from normal software: a bug here isn't a broken feature, it's a misstated balance, a failed audit, or a regulatory inquiry. I've spent a decade building exactly these systems — at Coinbase (crypto) and Nium (cross-border payments) — where "it mostly works" isn't good enough, because the books have to be honest down to the cent.
Settlement is the gap between a transaction happening and the money being final and reconciled in your books. "T+30" means that took thirty days; "T+1" means one. At Coinbase I owned the ledger modernization roadmap that automated accounting across hundreds of fund flows and compressed that settlement timeline by an order of magnitude — with full SOX/SOC guardrails intact. It matters because slow settlement is a business cost, not just a technical one: capital sits unconfirmed, close takes longer, and risk hides in the gap. Faster settlement means the business sees the truth sooner.
A ledger is the source of truth — the one place where a balance is real. Everything else (dashboards, APIs, caches) is a projection of it, and the moment a projection becomes a second source of truth, you effectively have none. A reconciliation engine is the system that checks reality against that truth — matching what your bank, your processor, and your ledger each claim happened, and surfacing the mismatches. At Nium I built reconciliation in-house from zero: near-complete auto-reconciliation across hundreds of bank accounts and many currency corridors, replacing a six-figure vendor. The hard part isn't getting most of the way there — it's the last stretch, where an unexplained cent is a thread you pull until the system tells you the truth.
Aggressively — but with humans accountable. At Coinbase I led AI-first engineering in production, not pilot: the majority of engineering work authored via AI, generative agents for subledger creation, AI-drafted PRs for production alerts so engineers remediate instead of investigate. But the guardrails are the point. You don't let AI vibe-code the systems that move customer funds. Every change passes compliance guardrails, quality gates, and human review. AI writes the code; a human still signs the SOX attestation and answers to the auditor. The acceleration is real — the accountability stays human.
The hard part of this job was never the technology — it's holding two things together that usually pull apart. You need the rigor to run systems where a single wrong entry is a regulatory event. And you need the trust to get a team's honest best — because in regulated finance, the engineer who's afraid to say "I think I broke something" is the one who causes your next audit failure. Precision without trust gives you cold rooms where nobody admits a mistake; trust without precision gives you missed standards. Both have to be true at once. I run high-trust teams because the stakes are high, not despite them — fear hides problems, and in money movement a hidden problem is the expensive kind. AI changes the tools every quarter. It doesn't change that the systems are built by people, and people do their best work when they're trusted, not watched.