model trained to recognize voices and restful api
Jysan bank model for voice recognition.
Given:
wav conversation btw 2 men - customer and tech support.
extract customer’s voice
convert to embedding
compare embeddings with cosine similarity, if score >90% - same voice, same customer