At Samsung R&D Institute Bangalore, I build and operate ML-driven production systems for browser privacy, onboarding, and platform reliability at roughly 100M daily active users.
I architected a hybrid RAG knowledgebase and onboarding agent using vector search, BM25, and Docker, reducing new-joiner ramp-up time from one week to two days. I've also shipped URL tracking protection and bounce-tracking defenses across mobile and desktop, combining ML classification, rule-based filtering, and live-traffic features to address tracking patterns across major navigation paths.
Previously, as an Amazon SDE intern, I developed serverless SQS-based applications for MTurk task processing, including error handling and retries, with 100% code and branch coverage.
