Technology executive · Payments, digital assets & applied AI
I lead the technology behind Mastercard’s global settlement, treasury and risk platforms — $12 trillion a year. Before that I ran engineering for The Clearing House’s RTP and CHIPS networks, and led the Fedwire re-platform at the Federal Reserve Bank of New York.
Co-founder, All Day AI 2026 keynote · Getting everyone in an enterprise to be AI-native →
Leadership
Fifteen-plus years at the center of the financial system — the Federal Reserve, The Clearing House, Mastercard — modernizing platforms that move trillions of dollars a day, while bringing AI into places where it has to be explained to regulators.
I retire legacy platforms without betting the institution on a cutover weekend.
Where fiat, stablecoins and tokenized deposits meet — built inside regulated institutions, not around them.
Agentic AI is only useful in a money-moving environment if risk, security and examiners can sign off on it.
Engineering cultures that move fast under Federal Reserve supervision and partner scrutiny.
Track record
I work as directly with customers, central banks and partner institutions as I do with engineers.
Mastercard
Oct 2022 – presentMost senior technology executive for Mastercard’s global settlement, treasury and risk platforms, and for AI adoption across engineering.
The Clearing House
Jan 2020 – Oct 2022Head of engineering for the U.S. private sector’s core payment infrastructure — RTP, CHIPS and the APIs banks and fintechs connect through.
Federal Reserve Bank of New York
Jan 2010 – Dec 2019Founded and led Common Services — Enterprise Architecture, Enterprise Data Management and the shared engineering, infrastructure and DevOps function serving all 12 Federal Reserve Banks.
TRG
Feb 2007 – Dec 2010Boutique advisory firm for digital transformation and enterprise modernization at Fortune 100 clients in financial services and life sciences, including Cigna and Covance.
University of Glasgow — B.Sc., Computing Science
Community · Co-founder
A free, online, follow-the-sun conference for AI practitioners — by practitioners, for practitioners. No vendor pitches: every speaker has shipped what they talk about.
I co-founded All Day AI with a group of engineering leaders to give the people doing the work a place to share the patterns, post-mortems and hard-won lessons you can’t get from a vendor deck. It also keeps me close to what practitioners across dozens of organizations are actually shipping — and where they’re struggling — with AI.
My 2026 keynote · LEAD track
Moving a whole engineering organization — not just the early adopters — to AI-native ways of working, inside an enterprise where security, risk and regulators all have a say. Drawn from bringing GenAI into the software lifecycle at Mastercard.
2026 speakers came from Microsoft · Accenture · Capital One · CVS Health · Qualcomm · Autodesk · MITRE · Carnegie Mellon SEI · W.W. Grainger · Anaconda · OpenRouter · Acronis — plus a four-track hackathon. All Day AI returns October 7, 2027.
Built by me
Outside the day job I build real products end to end — product, code, data and infrastructure — with AI-assisted development. It keeps my judgment about what AI can and can’t do grounded in practice, not vendor decks.
Live products, not mockups. Most now run on infrastructure I designed and operate myself — separate databases, auth, storage and serverless functions per product, automatic HTTPS and nightly backups.
The reference site for HTTP 402 and machine-to-machine payments: protocol guides, a registry, a validator and an ecosystem map for stablecoin micropayments and AI-agent commerce.
x402info.com ↗
Open-source, dependency-free ISO 8583 codec in Python with a UK Faster Payments profile — ASCII and BCD/binary encodings, the way switches actually send it. Published on PyPI under MIT.
github.com/tarvitave/iso8583-fps ↗
A weekly newsletter and library of practical AI patterns, tools and playbooks for people who ship — with AI-assisted publishing and automated delivery.
ayrai.com ↗
Micro-tipping for creators: tips as small as 2¢ with no per-tip fees, using embedded wallets and a card on-ramp.
my2centz.com ↗
Analytics on 99,000+ race results back to 1956: runner lookup, head-to-head comparisons and split pacing up a 14,115 ft summit. I run the Ascent; I wanted better data.
pikespeaklab.com ↗A browser-based command center for running many Claude Code agents in parallel — orchestrator, isolated worktrees, release trains and a live Kanban.
Read the write-up →Thinking
On settlement, stablecoins and tokenized money, real-time rails — and the engineering organizations that build them.
Latest · March 22, 2026
If you’ve spent any serious time building with AI coding tools, you’ve probably hit the same wall I did. You open Claude Code to tackle a task. Then you realise you need something running in parallel — so you open…
Read the essay →Written in 2018–19 as chapters of a book on AI governance — the question of who decides what an algorithm is allowed to do. Lightly proofread; otherwise as originally written.
More of my writing — on stablecoin settlement, tokenized deposits and AI in regulated engineering — is on LinkedIn.
Follow on LinkedIn ↗I enjoy conversations about the future of settlement — stablecoins, tokenized deposits and real-time rails — and about bringing AI into engineering organizations that can’t afford to get it wrong.