An event-driven backtesting suite meant to simulate historical trading environments for the development of algorithmic trading strategies. The program draws inspiration from the equities backtester outlined in Successful Algorithmic Trading by Michael Halls-Moore and the Backtesting.py framework.

The program is written in C++ and uses Python scripts and a MySQL database to store historical securities data.

Sky Spectre is a chess engine developed in C++ and rendered using the Raylib library. The first version was created using a basic mailbox approach for board representation programmed with Python3 (here). The current version implements the following features:

• Bitboard based game-state representation
• Fast move generation using bit manipulations
• Min-max move search algorithm optimized with alpha-beta pruning

A machine learning model developed as a part of a research project with the goal of implementing neural networks to quickly identify strokes, as well as classify them into the correct sub-category. The project in its entirety required:

• Thorough data pruning and normalization using publicly available databases
• Construction of the neural network using Python3 and the Keras and Tensorflow APIs

A complete version of the final thesis may be found here.

Graphity is a graph visualization web application created for HackUMass XI using React.js and TypeScript. The website is targetted at students and developers alike to easier understand, create, and implement graphs and graph algorithms in their programs. Its features include

• User-friendly drag-and-drop interface for graph creation and manipulation
• Built-in code editor and IDE for operations on the user created graph
• JSON generation for user-created graph for personal use

An android app designed for small business owners seeking centralized management. Developed with Android Studio and Kotlin, EmployBee allows for seamless communication between an employer and employees, including features such as:

• Customizable protocol search engine powered by an SQLite database as an alternative to employee handbooks
• Automatically refreshing and updateable shift-specific tasks stored in SQLite database

A sentiment analysis trading strategy created to test the viability of making trades based off of data collected from tweets. Tweets were gathered and cleaned before being used to create a model that would predict the sentiment of the market. The model would then make trades based off of the sentiment. The project was back tested over a year and shows promise, resulting in 38% returns, outperforming the S&P 500 by 2.53 times.

A maze generation and solver program created using common graph traversal algorithms. The maze itself is originally generated (as seen in the demo video) using depth-first search and recursive backtracking. The program is then capable of traversing the maze using a variety of algorithms present in graph theory.