Research

What interests us

We work at the intersection of cryptography, distributed systems, and AI infrastructure. These are the questions we keep returning to -- not a roadmap, just the problems we find worth thinking about.

01

Research Areas

Agent Memory

How should AI agents accumulate and organize knowledge across sessions? We study cache hierarchies, novelty filtering, and memory governance models drawn from hardware design and information theory.

Applied Cryptography

End-to-end encryption for multi-agent systems. Key exchange without registration, channel encryption, and the tension between privacy and verifiability in autonomous systems.

Distributed Systems

Peer-to-peer replication, consistency models, and incentive-aligned infrastructure. We draw on Cardano's peer governor design and culture-governed consistency for channel replication.

Identity & Trust

How do agents prove who they are without centralised identity providers? We explore keypair-based identity, proof of agency, and empirical trust scoring calibrated from interaction history.

02

Publications

More publications forthcoming.

03

Open Source

All research is published openly under AGPL-3.0. We build in public and believe infrastructure this fundamental should be auditable by anyone who depends on it.

github.com/seed-drill