Fnu Hridyanshu is an undergraduate researcher at UC San Diego and a member of the HXI Lab. His research spans human-centered AI, human-computer interaction, 3D design, computer graphics, spatial computing, digital twins, multimodal systems, and simulation. At HXI, he works on the CLARO project, exploring human-centered interactions with intelligent systems.
His research focuses on computational systems that model and interact with both the physical world and human behavior. He is particularly interested in combining data-driven models with physics-based simulation to construct digital twins, while also using large-scale data to model human, social, and socio-emotional behavior. Across these areas, he is interested in reducing the gap between human intent and the complexity of the computational systems required to realize it.
Hridyanshu created ResonantLoom, an object-centric generative sound-design system that enables AI-generated video audio to be edited at the level of individual entities and sounds. The project appeared at ACM CHI 2026 and ACM SIGGRAPH 2026, where it received Third Place in the Undergraduate Student Research Competition.
His earlier work includes UmeedVR, a conversational AI and virtual-reality platform for autistic and neurodivergent users. UmeedVR was subsequently acquired, making Hridyanshu one of the youngest people to have an AI product acquired. Developed before the public release of ChatGPT, the system combined immersive 3D environments, conversational interaction, and natural-language processing for communication and therapy-oriented experiences.
3D design has been a recurring foundation of his work across CAD, computer graphics, virtual environments, spatial interfaces, and engineering simulation. As part of the winning team of the NASA Lunar Loo Jr. Challenge, he worked on system layout, CAD, mechanisms, fluid dynamics, safety, packaging, and engineering under strict mass, volume, and power constraints. NASA engineer Michael A. Interbartolo III specifically highlighted his CAD and mechanisms expertise and responsibility for system layout. Read the NASA recommendation letter.
ResonantLoom: From Prompting to Repair in Object-Centric Generative Sound Design
CHI EA ‘26: Extended Abstracts of the 2026 CHI Conference on Human Factors in Computing Systems
ResonantLoom: Neural Acoustic Rendering for Editable Multi-Entity Video-to-Audio Synthesis
SIGGRAPH Posters ‘26: Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Posters
Thermodynamically Consistent Neural Models of Protein Folding Kinetics
BCB ‘25: Proceedings of the 16th ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics
B.S. in Mathematics-Computer Science, 2027
University of California San Diego, San Diego, CA, USA
Minor in Cognitive Science, 2027
University of California San Diego, San Diego, CA, USA