Project Tusk
A full-stack AI platform for ElephantVoices that combines custom audio processing, machine learning, LLM-powered reasoning, and real-time interfaces to help researchers analyze elephant vocalizations.
The system removes noise from audio, extracts acoustic features and MFCC voice fingerprints, classifies elephant call types, and turns the results into readable research summaries.