Projects

Scientific Claim–Citation Verification Engine

A research-focused system that traces scientific claims back to their cited sources to evaluate alignment between assertion and evidence. It extracts explicit claims, maps them to references, and retrieves relevant source context for comparison. The engine classifies support levels (e.g., supported, misaligned, contradicted) with uncertainty-aware scoring. Designed as a verification assistant rather than a credibility judge, it prioritizes interpretability over black-box prediction. The project explores how AI can strengthen rigor in high-density citation environments such as AI research.