Multilingual Text and Fintech Transaction Labeling for AI Model Training
Participated in large-scale data labeling and model evaluation for multilingual text and financial transaction data. Labeled and validated entities, intents, and relationships across thousands of samples to train natural language understanding models for real-time categorization of financial transactions. Designed and reviewed synthetic text prompts and responses for fine-tuning large language models (LLMs). Ensured data quality through multi-tier validation and compliance with project accuracy thresholds (>98%).