AI & ML interests

HSRS is dedicated to advancing persistent, emergent 3D spatial AI runtimes such as perception bubbles and PerceptiveNeuralOS that deliver robust spatial awareness and long-term memory for embodied intelligence. Our mathematically grounded approach enables ultra-dense, self-evolving computation in voxel grids, seamlessly integrated with vector-based retrieval-augmented generation and multimodal interfaces (images, prompts, and real-time inputs). We focus on practical outcomes including real-time 3D object generation, atomic-scale simulations, shared avatar environments, and applications in AR/VR, robotics, ethical oversight, and education. Near-term priorities include rapid development of profitable prototypes (e.g., webcam-driven AR tools and simulation APIs) using established frameworks, while long-term objectives encompass revolutionary advances through analog hardware integration, verifiable emergent behaviors, and scalable persistent intelligence. We actively seek collaborations in analog computing to accelerate efficient, energy-optimal deployments.