The Company

Deep curiosity, guided by rigour.

Ryvane is a research and training company building the toolchains, curriculum, and operator practice the AI security field actually uses.

Tomorrow's threats are already trained against today's AI systems. Our mission is to make offensive AI security a discipline practitioners can train into.

At Ryvane, we believe that security is not static. It's a system in motion. Our mission is to help the world's most ambitious teams turn live AI security research into engineering practice, designing the toolchains, training programs, and audit playbooks that move adversarial work into shipping defenses.

We don't just consult. We operationalize. Ryvane works side by side with engineers, security teams, and operators to navigate complexity with clarity, and to architect systems that scale discovery into impact.

This is what drives us: the pursuit of questions that matter and the precision to act on what we find. We bring deep expertise, radical focus, and systems-level thinking to every challenge.

Arun NairCo-Founder

Our team

Andre Brosio

Andre Brosio

Lead AI Researcher

Lian Owen

Lian Owen

AI Researcher

Ajith George

Ajith George

Lead Dev

Arun Nair

Arun Nair

Co-founder

Shaping AI security in production

Real-world adversarial research turned into the tools, training, and audits security teams actually use.

23+Production engagements shipped

120+Operators trained through Academy

9Open-source toolkits maintained

Trusted by security and AI platform teams across four continents.

Continuum.aiEisner & SterlingGIVONNIGridpoint

Journal

MCP

Model Context Protocol

6/12/25
AI Security

The case for offensive-only training cohorts

6/2/25
AI Security

Three production agents, one sandbox escape, same week

6/2/25
Research

Memory poisoning in persistent agents

6/2/25
Research

Your retrieval pipeline is a security boundary

6/2/25
Open Journal