Who am I
It’s an ordinary Friday afternoon. I’m sitting in a Starbucks in Zurich, focused on my next project, when I connect to a public Wi-Fi network to send a couple of emails.
A few days later, something feels off. I start receiving messages—payments and registrations I never requested. At first, I dismiss them as spam. Then it becomes clear: someone has accessed my Gmail account.
[ me ] --- wifi --- [ attacker ]
\
\--> [ my data ]
What surprised me most was the lack of any warning—no notification, no sign of suspicious activity.
That moment sparked my curiosity about how our digital lives can be protected—and how easily trust can be broken. This led me to work as a researcher, contributing to projects focused on protecting photographers’ rights [startup website] and exploring decentralized alternatives to platforms like Airbnb [github].
Today, I work at the intersection of security and artificial intelligence, helping people build AI agents while staying grounded in the values that started it all: securing businesses and protecting people’s digital lives.
Next project(s)
AI agents are improving rapidly. However, they are usually designed as isolated systems, and they are built for a specific purpose.1
The direction I’m exploring is to design an AI agent orchestrator—a system where agents are not standalone units, but parts of a coordinated architecture. It consists of:
- AI agent skills. [github]
- Delegation framework. Research draft [hackmd.io]