At InstaHire, my capstone project, I built an AI resume screening engine that ranks candidates against job descriptions using skills, experience, and keyword relevance.
I also integrated Elasticsearch and Redis to support fuzzy job search and caching, reducing average query response time and improving scalability under concurrent load.
On Argument Cartographer, I developed a multi-agent system that extracts claims, counterarguments, and evidence from news articles. It improved extraction accuracy by 21% over traditional prompting across 48 real-world articles and generated explainable argument maps with source citations.
I developed SmartSpend, a personal finance platform with OCR receipt extraction and expense categorization, and QueueVision, a computer-vision pipeline for monitoring queues from live CCTV footage. My research comparing prompting, ReAct, and multi-agent approaches was presented at ICCUBEA 2026.
