• Working as a Software Developer focused on observability and reliability for large-scale distributed systems across on-prem and cloud environments, building Python-based telemetry pipelines that transformed infrastructure and service metrics into production dashboards and alerts using Grafana, Splunk, OpenTelemetry, and Wavefront.

• Contributed to the training, evaluation, and refinement of Large Language Models (LLMs) across 5+ high-impact projects, improving model accuracy and reliability while performing in-depth safety and bias assessments to ensure ethical, production-ready outputs.

Personalized travel share links with interactive map visualization

A full-stack application that lets you create personalized share links showcasing all the places you've visited. Interactive map visualization powered by Mapbox displays your travel locations, while PostgreSQL stores your travel data securely. Upload and manage travel memories with AWS S3 bucket integration.

A full-stack injury-prevention platform

Built a full-stack web platform that analyzes soccer injury patterns to support ACL/MCL tear prevention, using Next.js, Node.js, Express, and PostgreSQL. The platform applies predictive analytics with TensorFlow.js and OpenCV to evaluate over 500 annotated movement samples, identify high-risk motions, and recommend targeted warm-ups and training drills to help reduce injury risk.

NLP-Powered Financial Search Tool

Built an NLP-powered financial search tool that interprets free-text user queries and maps them to relevant stocks and ETFs, improving discoverability beyond traditional ticker-based search, and evaluated search accuracy using real user feedback, achieving approximately 85–90% relevance accuracy based on participant rankings of returned results.