Audio Model Trainer — Contract Role
In this project, I generated spoken descriptions for images to train audio models for natural language understanding. Recordings were created using a specialized internal platform, with careful attention to consistency, clarity, and dataset requirements. I worked closely with AI researchers and QA to ensure that recordings and labels met standards for dataset quality. • Created natural-sounding spoken descriptions for images. • Delivered consistently high-quality audio free from distortion. • Adhered to stringent linguistic and data collection protocols. • Supported iterative dataset improvements with AI/QA teams.