AI-Assisted Data Labeling and Evaluation (Personal/Academic Projects)
Transformed unstructured business emails and documents into structured JSON outputs for downstream processing. Defined annotation categories and criteria to support data labeling QA and AI-assisted extraction validation. Reviewed and corrected model-generated outputs for consistency, ambiguity resolution, and schema compliance. • Developed clear annotation guidelines covering categories such as order emails, quotation inquiries, and follow-up status. • Supported dataset cleanup and quality checks for AI-generated content. • Iterated on prompts and extraction templates to improve classification accuracy. • Ensured technical, bilingual, and business-domain content was properly structured and validated.