Research Assistant, Economic Entity Recognition
I developed and fine-tuned Transformer-based Named Entity Recognition (TNER) models, specializing in financial documents. My work involved annotating over 50 financial documents using the IOB format to extract economic entities, and aggregating labels from multiple annotators to ensure precision. We strictly adhered to a comprehensive annotation guide and used Label Studio for the annotation process. • Created training datasets by labeling entities such as organizations, individuals, and financial instruments. • Collaborated with a team to resolve annotation disagreements and standardize labeling practices. • Evaluated model performance with labeled data to improve NER accuracy. • Used Label Studio extensively for document annotation and quality assurance.