01 // The Problem Statement
Traditional phishing filters lack explainability, leaving users unaware of why specific emails or links were flagged as threats.
02 // The Technical Approach
Built an Explainable AI (XAI) phishing detector providing feature attribution, IP geolocation, and threat visualization dashboards.
03 // Key Failures & Challenges
Balancing model accuracy with real-time prediction speeds for interactive REST API evaluation.
04 // Engineering Insights & Learnings
Implemented SHAP/LIME explainability techniques with Scikit-learn classification models and Flask backend APIs.
05 // Future Scaling & Improvements
Deploy browser extension for real-time URL inspection.