At Samsung R&D Institute India (SRI-Delhi), I resolved more than 120 Kona defects in a multithreaded C++ codebase and submitted changelists through Perforce. I also owned the AVControl module and handled its deployment lifecycle.
I engineered a generative AI log analyzer that routes Kona tickets through a multi-agent LangGraph workflow. It uses a FAISS-backed RAG pipeline to match historical defects, while an MCP server retrieves relevant code selectively; my LLM-based crash investigation work reduced system diagnosis time by 90%.
Earlier, as a Summer Intern at LTIMindtree, I extracted company data into MySQL, developed a Flask dashboard, and contributed to an OCR-based application. My projects include a generative AI personal assistant, an offline-first finance tracker, and a movie and series platform.
