systems Self-Evolving Agent Hierarchical multi-agent system that runs the threat-to-patch loop autonomously. Multi-Agent Research Harness Coordination layer for concurrent autonomous agents working on the same codebase. Co-Scientist AI agent that drives autonomous research workflows end-to-end. Super Co-Alignment End-to-end alignment for multi-modal foundation models. Digital Identity Defense Tech lead for Amazon Rekognition digital identity at billion-level annual scale. GPU-Accelerated FHE Inference Privacy-preserving inference and similarity search at production speed. current Continual Learning for FMs Continual learning that drops the disjoint-task assumption. Hybrid Attention/Recurrent FM Linear-time recurrent layers (GKA) for long-context agentic workloads. Hardware-Aware Distillation Training methods for analog in-memory computing inference. earlier Omni-Modal Trust & Verification Detecting tampered media across image, video, and audio with explainable reasoning. Liveness Detection Production face liveness with temporal consistency, behind billion-scale identity checks. Deepfake Detection Self-consistency and language guidance for generalizable deepfake detection. Billion-Scale Face Recognition Pose- and occlusion-invariant face recognition behind production identity at scale.