Voice Data Annotator (Cosmos Project)
As a Voice Data Annotator for the Cosmos Project with OneForma, I performed voice labeling and interaction segmentation to support AI voice assistants and large language models. The work involved analyzing spoken South African English conversations for intent, context, and interaction points. I maintained high accuracy while working with extensive audio datasets to enhance voice recognition systems. • Labeled and segmented voice data for AI assistant training • Identified conversation intent and context • Applied consistent standards to audio annotations • Contributed to NLP and LLM improvements