Prithviraj Manivannan

Conducted a comprehensive literature review on ML driven financial fraud detection to identify critical gaps in model generalizability and data bias. Proposed a scalable, ML pipeline for fraud detection integrating feature engineering, and explainability techniques — with emphasis on reducing false positives while maintaining high recall.

Created a tool to automatically extract information of almost all employees of a target company on LinkedIn. Reduced lead collection times from a week to 2 days. Learned and deployed techniques to combat anti-scraping detection such as human cursor emulation, driver manipulation, etc.