Kawaii - Audio Annotation and Speaker Identification, United States

We’re looking for native speakers to help train advanced speech recognition models by labeling conversational interactions. You’ll listen to audio recordings and apply precise, time-stamped labels to identify speakers and mark key moments in human-AI conversations. This is a great opportunity to contribute to cutting-edge AI development while working on your own schedule from anywhere. 

The goal is to accurately identify and label conversational events within recorded audio files to help train and evaluate advanced speech recognition models. Each recording features a User interacting with Siri (or another voice assistant). Your precise labels will help the model learn to recognize and understand natural speech patterns, regional accents, and assistant responses. 

• Native-level fluency in one of our target languages: Vietnamese, Malay, Norwegian, Thai, Turkish, Hebrew, or Traditional Chinese 
• Based in the target countries

• Strong listening comprehension and familiarity with regional speech patterns and accents 

• Ability to pass OneForma Certification and KYC verification 

• Excellent attention to detail and ability to follow structured linguistic guidelines 

• Ability to distinguish multiple speakers and track conversational flow 

• Access to a reliable computer, stable internet connection, and an iPhone 

• Comfortable working with web-based annotation tools 

• Previous experience in annotation, transcription, or data labeling is a plus