Cosmos – Voice Assistance Interaction Segmentation

Description: 

We are looking for motivated contributors to join the Cosmos Project as Voice Assistant Interaction Segmentation and Labeling & Transcription Specialists.

Contributors will listen to long-form audio recordings between users and voice assistants and segment them into distinct, meaningful user interactions. This ensures that each voice assistant-user exchange is accurately identified and contextually complete — helping enhance conversational AI systems and voice recognition models. 

Contributors will also label speaker turns and accurately transcribe user speech, following detailed annotation guidelines and localized trigger phrases specific to the target language. Their work will ensure that each user–Voice Assistant exchange is accurately identified, contextually complete, and linguistically precise, helping improve conversational AI and voice recognition models. 

Purpose: 

This project...