Building computable commerce.
Personal shopping agents are the first wave of what is coming: Automated economic decision making at scale. Agents can only go as fast as they can check the rules are followed and their counterparty is trustworthy. That is the emerging bottleneck. I'm interested in building tools and infrastructure that alleviate this bottleneck.
Code verification and cryptography are the building blocks that can make the agentic economy trustworthy. I run the Lean in Munich meetup, and was a guest lecturer in cryptography at Technical University of Munich.
I've built an MCP server that generates compliant invoices via chat, an AI tool that calculates beneficial ownership from private documents in minutes and a simple DNS-based merchant verification framework for x402 resource servers. I use cryptography and formal languages to make sure that every decision remains explainable, auditable and within constraints.
This world requires builders that can interface with governments and understand the liabilities that can arise in moving real money. I've done both.
Background
Director General, Bavarian Ministry of Digital Affairs: Cloud and AI strategy · EU digital regulation · international affairs
Derivs Trader & Quant: JPMorgan, Highbridge Capital
Lex columnist, Financial Times: Analyzing the intersection of regulation, geopolitics and business
Academics: PhD Applied Mathematics, Brown · Guest Scientist, TUM
Email me
See what I'm building ↓
How I can help you
I fund myself by working with investors and companies that are advancing the same frontier:
Formal verification
For investment teams
What's real, who's serious, and what to ask a founder.
I brief you on my two theses: 1) If it's not worth verifying, then it's not worth engineering. 2) Legal services can scale and automate through formally verified code.
briefing + memo.
Email about a briefing →
Trust design review
For teams building harnesses that want to harden them with verification and cryptography
How to enforce deterministic constraints and create auditable evidence traces
Trusting agents is a cryptography problem: identity, access, delegation, and auditability.
Half-day review + memo.
Email about a review →
Frontier regulation in AI, compute and commerce
For CFOs, policy teams, compliance leads and investors
What emerging policy areas are critical for my business and investments?
The shape of future commerce is already written in the AI Act, GENIUS Act, MiCA, and AML law. I bring the global picture together for you.
briefing, workshop, memo
Email about an engagement →
What I'm building
Working prototype
Invoices from chat. Auditable, and explainable, generated from code verified against the legal rules a tax adviser checks.
See how it works →
Ownership (UBO) checker
In development
Determining a UBO requires following control relationships that emerge from private documents, not public registers. I'm building tools that provably follow the applicable rules, without revealing private information.
Working prototype
A key in your DNS record binds a wallet to your domain, so an agent can check before paying.
See how it works →
Where to find me
If you're there, email me and we'll find twenty minutes.