Audio Model Trainer, Mercor.io Corporation (Remote)
Recorded short audio clips describing visual content to support building and refinement of datasets for multimodal AI systems. Performed dataset contribution work by translating what was visually observed into corresponding audio utterances for model training. Ensured recordings align with intended dataset requirements for downstream AI training. • Captured audio descriptions from visual content • Supported multimodal dataset development • Contributed to iterative refinement of training data • Worked remotely for the labeling/training pipeline