Multimodal AI Dataset Creation & Annotation for NLP and Computer Vision
Developed and annotated high-quality datasets across text, image, and audio modalities for AI model training. Designed prompt-response pairs for supervised fine-tuning (SFT) to improve conversational AI performance. Labeled image datasets using bounding boxes and segmentation techniques for object detection tasks. Transcribed and annotated audio data for speech recognition and classification models. Performed Named Entity Recognition (NER) and text classification for NLP systems. Conducted quality assurance to ensure consistency, accuracy, and adherence to annotation guidelines. Contributed to improving model performance by providing structured feedback and error analysis.