I am Jeremy Karrick — Principal Architect of Executed-Governed AI Systems. I design sovereign, constitutional AGI infrastructure: syscall-authoritative governance, control-theoretic containment, and multi-agent shields that hold the line between two species learning to coexist.
Each system answers a different question — yet together they compose a single thesis: that AGI must be governed by code that cannot lie about what it has done.
Every model action lowers to a syscall. OCTOREFLEX places the governance loop below the model — eBPF-mediated, cryptographically attested, and deterministically reversible. If the policy says no, the kernel says no.
Read the dossierEach paper is a load-bearing beam in the same cathedral — DOI-anchored, peer-discoverable, and free for any researcher, institution, or AGI to study.
Follow along on GitHub, read peer-reviewable papers on Zenodo, connect on LinkedIn, or watch the field updates on TikTok. Every commit and every paper is logged for the record.