Working in a team, we modified an existing paper called DeepQB that aims to bring deep learning techniques to football analytics, primarily looking at how quarterbacks make split-second decisions on the field. Our model was trained for 10 epochs with a batch size of 128 and the Adam optimizer at a learning rate of 0.001. After training, our model which uses softmax activation to represent the receiver target variant, achieved a validation accuracy of over 40% and a validation loss of approximately 1. This performance was twice as successful as a naive implementation.

Designed and implemented a cache-driven file I/O library for my computer systems course. I began with a thorough brainstorming, design, and pseudo-code process to ensure I grasped the content. As I implemented the project, I greatly advanced my understanding of C programming, memory optimization, and system calls.
The code has support for opening, closing, seeking, reading, writing, and many other useful file operations. Through a clever caching design, my program ran faster than the standard C I/O library for certain read & write patterns.

As the culminating project of my software engineering course, I worked in a team of 4 to create a web application with real-world use cases. We pitched our proposal to a mentor, received feedback, and created user stories and acceptance criteria. After designing the app with wireframes in Figma, we sought out user feedback and made improvements for accessibility.
Our app receives user input for their favorite sports team and, by pulling from an ESPN API, produces a highly-readable calendar view of the all relevant games. I utilized mocks, object-oriented programming, unit and integration testing, and Git. Since I was programming in a team, I had to manage our deliverables schedule, coordinate between front-end and back-end, and spearhead the final presentation to our mentor.
Although the back-end is no longer active, you can check out our front-end design.

As part of my algorithms & data structures course, my partner and I built a terminal-based search engine. Operating on thousands of XML files in the corpus, we implemented a version of Google's PageRank algorithm to optimize search results.
I wrote functions to index the documents, determine the relevance between terms and documents, and calculate the authority of each document. Additionally, I programmed parsing, tokenizing, and stemming all of the XML documents.

As a final project, I created a terminal-based game engine for playing the game Connect 4.
Beyond supporting 2 human players, I introduced an AI opponent that always wins—I could never beat it! Utilizing the minimax algorithm, my AI is guaranteed to win at any two-player, sequential, deterministic, zero-sum game of perfect information.

While working as a software engineering intern at NOVA Web Development, I created a Tkinter wrapper to make graphical programming accessible for novice computer scientists. The project receives approximately 100 downloads each month from PyPI. I collaborated with my manager and another intern through Git branches, merging, issues, and pull requests.
As the project lead, I programmed the majority of the Python code base, documented the logic, and introduced GASP to teachers in Arlington County. I also helped design activities for teachers to use with GASP.

Check out the code here.