AI-Powered Content Moderation Dataset Creation
I led the creation and annotation of a 10,000+ record dataset of social media posts focused on content moderation for a local startup. I performed multi-label classification and sentiment analysis using LabelStudio, with quality control measures to achieve over 98% annotation accuracy. I developed custom scripts to automate collection and preprocessing tasks for efficient workflow. • Labeled large volumes of social media text data for content safety. • Applied multi-label classification and sentiment annotation for nuanced categories. • Utilized LabelStudio extensively for labeling and quality reviews. • Ensured dataset usability for training an AI-powered moderation model.