Independent AI safety researcher building SEVERANT, a formally verified, hardware-constrained AI architecture where unsafe behaviour is physically impossible by design. Working at the intersection of formal verification, hardware architecture, and AI safety. Sole researcher, no institutional affiliation.