AI Data trainer
In this project, I worked on training data for large language models with a focus on audio tasks. My responsibilities included evaluating and rating audio outputs for accuracy and naturalness, collecting and organizing audio data, writing prompts and responses for supervised fine-tuning (SFT), and contributing original audio recordings to expand the dataset. I consistently applied detailed guidelines to ensure quality, balanced coverage, and alignment with project objectives. I used internal and proprietary labeling tools to carry out these tasks, maintaining accuracy and efficiency throughout. The project emphasized strict quality assurance, with regular reviews and feedback cycles to refine both data and model performance. My contributions helped improve the model’s ability to understand and generate human-like responses across different contexts.