- Created a tool to detect resume fraud and suspicious endorsements in recommendation letters by identifying falsified claims.
- Created features such as Loyalty Index , Redundancy Score (repeated endorsements), Gap Ratio (employment gaps) , and Exaggeration Score (frequency of praise words) to assess resume and recommendation letter authenticity.
- Used NetworkX for circular endorsement detection and the Leiden algorithm for clustering suspicious patterns.
- Identified fraudulent resumes and patterns, equipping HR with a dashboard for clear insights and informed decisions.