// signature work
Original research and products: Vani (contrastive acoustic voice profiling), a real-time Space Weather Alert ML model, and MLAF — a patent-pending framework for multimodal language acquisition.
Contrastive Acoustic Voice Profiling System
A speaker-recognition architecture using contrastive learning over acoustic embeddings — the namesake project behind Vaaani's mission of giving language back to machines, faithfully.
Demo on request →Real-time ML for solar & geomagnetic activity
An end-to-end pipeline ingesting heliospheric data streams and predicting space weather events that affect satellites, communications and grid stability.
Demo on request →Multimodal Language Acquisition Framework
An original framework for modeling how humans acquire language across modalities — currently under provisional patent. Brings linguistic theory into hands-on ML systems.
Demo on request →