About Jeremy

Jeremy Rose

Founder of Odins LLC

I'm a security and systems professional with a background spanning COMSEC, cryptographic key management, federal systems, public safety technology, operational risk, and now AI security and governance.

My work today focuses heavily on the problems emerging as AI systems become more capable and autonomous: prompt injection, adversarial behavior, capability and authority boundaries, containment, recursive risk, threat modeling, and how organizations maintain control when AI can reason and act at machine speed.

I'm also the founder of Odins LLC and the creator of Odins Quantitative Governance Framework (OQGF), where I research and develop approaches for governing AI systems without placing the model itself in the trust path. A principle that guides much of my work is simple: the model may propose, but deterministic controls should govern consequential actions.

My working style is grounded in evidence. I document constantly as part of the process. Notes give me continuity across complex work, help make results reproducible, and give me a record to return to when something fails. When a problem appears, I want to be able to trace what changed, identify the assumption or gap that produced it, correct it, and turn what I learned into a stronger process or regression test.

That same mindset shapes how I approach AI security: understand the system, challenge its assumptions, find where the boundaries fail, preserve the evidence, and make the next iteration harder to break.

I'm particularly interested in AI security, AI red teaming, adversarial ML, autonomous agent security, AI governance, model risk, and the engineering controls required to deploy increasingly capable AI responsibly.