Building
01
Akaeon
Cryptographic infrastructure for digital authorship and AI-training consent. The parent of everything below.
02
Stelais
A receipt for making something. Timestamped, verifiable proof of authorship for creators — a permanent record of what you made and when.
03
Omnia
A language app for functional travel competency, not fluency. French and Arabic first. Paper-and-ink interface.
Premise
The internet has no built-in concept of “I made this first.”
Research
ORCID
01
Vision-Encoder Behavioral Fingerprints of Image-to-Image Generative Models
A training-paradigm-driven taxonomy of six commercial APIs. Edit-trained and text-to-image-base models separate into two behavioral bands.
02
Open corpus and supplementary materials
Datasets and supplementary bundle for the fingerprinting paper. Open deposit with a restricted companion.
About
I make records — infrastructure that lets a person prove they made something, and research into the machines that learn from it.
I’m the founder of Akaeon and Stelais. The work sits where cryptography, machine learning and creator rights meet: verifiable records of authorship, anchored permanently, that survive the pipelines training the next generation of models.
Alongside the companies I do independent research on AI model attribution and behavior, published on arXiv. Before any of this I was a professional snowboarder. I serve on the board of directors at the University of Vermont.
I write on Substack. Fewer, better decisions. Let the work speak.