AI agents and dynamic identities move faster, touch more data, and create more sprawl than static controls were designed for. Aurva was founded to answer one question: when does legitimate access become inappropriate at runtime?
Built and led AI systems at Meta. Understands how production models behave with sensitive data at scale.
Built security infrastructure at Meta focused on access linked to real identities and context.
Systems architect shaping the data plane and eBPF-based observation layer.
Built with support from the investors and operators who shaped modern AI, security, and cloud infrastructure.
World-Class Investors



Industry Veterans








We're a small team solving a hard problem. If runtime security for AI interests you, reach out.