Below are a few of my more recent projects that I am particularly proud of. If you want to browse the source code for any of my projects, they can all be found on my GitHub here.
The goal of this project was to create a service that analyzes network traffic from different devices. Initially inspired by a curiosity about smart home device traffic, it evolved into a tool that visualizes network activity for any device of interest. Users can upload a PCAP (packet capture) file and specify an IP or MAC address to analyze, then view visualizations of traffic volume over time, communication partners, and geographic destinations. The application is live at https://pcaptracker.site, or users can run it locally via Docker by cloning the GitHub repo.
Key Features:
A recreation of the Git version control software built from scratch in Java using Gradle for consistent and quick builds. I developed this project to explore how Git operates internally and interacts with the file system.
Key Features:
A major challenge in medical machine learning is acquiring large, high-quality datasets for training models. To address this, my team and I developed a few-shot learning model that requires significantly fewer images to train while maintaining high accuracy. Our model needed only ~50 images to achieve ~88% accuracy, compared to conventional models requiring thousands.
Key Features:
Developed an Android application in a team of six using Agile development practices. The app allows users to add friends via unique codes and track their real-time locations on a dynamic compass UI. Friend locations update continuously as they move throughout the day, with data synchronized via a cloud-hosted AWS SimpleDB server.
Key Features: