Lithuania
I build the perception layer modern AI systems are missing real-time human physiological and cognitive state extracted from a standard RGB camera, fully on-device. High-stakes environments fail for the same reason: systems see the world, but not the human operating inside it. Fatigue, overload, stress collapse, uncertainty these shape outcomes long before performance drops. Today’s AI stacks have no visibility into any of it. My work focuses on closing that gap. A five-layer ontological architecture that interprets human state in real time, runs in contested and air-gapped environments, and is built to integrate into defense, clinical, and dual-use workflows where stability under pressure matters. I build systems that keep human–AI coordination coherent when conditions degrade.
Ontological infrastructure for high‑stakes AI. Neurosymbolic engine, rPPG biometrics, coherence scoring, and drift detection.