Nicholas Limarsha

A monorepo of five quantitative-finance tools, each built to do one job well with tests, benchmarks, and a runnable demo: an event-driven backtester, a real-time market-data pipeline, an options pricer (Black-Scholes, binomial trees, Greeks, implied vol), a C++ limit-order-book matching engine, and an MPT portfolio optimizer (efficient frontier, VaR/CVaR). Shared CI, Docker, and a dev container.

A user interface for a search engine that incorporates Wikipedia searching. Built with React and utilizing Wikipedia's API through Axios for seamless data fetching.

A Spring Boot application that includes user authentication with JWT-based security for user sign up and login with email verification. Features email verification, JWT token authentication, and comprehensive error handling.

Recommends workout videos by matching a user's uploaded exercise clips against a database — comparing movement sequences and similarities from body-keypoint extraction, then surfacing relevant recommendations.

An image-based ML process to identify & annotate whether people are wearing masks. Using a dataset from Kaggle with images of people wearing masks, not wearing masks, or wearing masks incorrectly, along with PASCAL VOC format annotations.