Financial Transaction Data Annotation for Fraud Detection Model
Text
ClassificationWorked on a data annotation project focused on financial transaction records for fraud detection model development. Responsibilities included: Annotating over 15,000 transaction records including payment logs, merchant data, and user behavior signals Labeling transactions as normal, suspicious, or fraudulent based on predefined risk guidelines Applying consistent annotation rules to handle ambiguous or edge-case transactions Performing multi-round quality checks and resolving inconsistencies, achieving over 96% labeling accuracy The dataset was used to support machine learning model training for fraud detection and risk scoring systems.
2025 - 2026