About

Building infrastructure for trusted visual media.

Sigillum is focused on long-term infrastructure for authenticity, provenance, and protection in an environment where generative systems can reproduce and recombine any image or video at scale.

Mandate

A foundational layer for visual trust.

The integrity of visual media is no longer a property of the file. It is a property of the system around it — the protections embedded into the asset, the provenance bound to its identity, and the verification surface available to anyone who encounters it.

Sigillum is building that system. Our work spans adversarial machine learning, signal processing, cryptography, and the operational realities of large-scale media pipelines.

We approach the problem as infrastructure: durable, well-specified, independently auditable, and built to last beyond any single generation of generative models.

Principles

What guides our work.

Research-driven

Decisions grounded in measurement, not narrative.

Privacy-conscious

Protection without surveillance. Data minimization by default.

Technically rigorous

Specifications, threat models, and verifiable claims.

Long-term thinking

Built for a decade-long horizon, not a product cycle.

Disciplines

Where our work sits.

Adversarial ML
Signal processing
Cryptography
Computer vision
Codec engineering
Trust & safety
Forensics
Standards work
Distributed systems