MSc Data Science & AI at TU Delft, Netherlands.
I'm building TrustChain, an open-source trust layer for AI agents. The idea: every agent-to-agent interaction creates a bilateral cryptographic record signed by both parties. Trust scores come from real interaction history, computed via max-flow network analysis so fake identities can't game the system.
It's based on the TrustChain protocol from TU Delft (Otte, de Vos, Pouwelse). I extended it with NetFlow trust computation and filed an IETF draft. Also submitted a response to the NIST CAISI RFI on AI agent security.
pip install trustchain-py
from trustchain import with_trust
@with_trust(name="my-agent")
def main():
... # all HTTP calls now go through the trust proxy| Repo | What |
|---|---|
| trustchain | Rust core, QUIC P2P, HTTP proxy, dashboard, MCP server |
| trustchain-py | Python SDK, @with_trust decorator, CLI, framework integrations |
| trustchain-js | TypeScript SDK, OpenClaw plugin |
| trustchain-agent-os | Adapters for 12 agent frameworks, MCP gateway |
- hackaton-dublin - Skyly: 3D data center feasibility estimator for Earth/Moon/Mars. React, Three.js, Rust, Solana. HackEurope Dublin 2026
- haxbet - Stock market for hackathon ideas. Hit 130 concurrent users, patched live to fix botting. HackEurope Dublin 2026
- partydrinks - OutDrink your friends. Live drink stats for parties
- claude-usage-widget - Track your Claude API usage
- Spotify-Ranker - Rank Spotify albums against each other
- Computer-Graphics-project - C++ renderer, built for uni
- Chess - Chess game in NodeJS