Projects

A curated selection of recent work, explorations, and contributions

Featured

FundusVision: Automated Retinal Disease Detection

Led 5-person team developing automated retinal disease detection system for CS 7641 Machine Learning. Implemented three approaches: U-Net vessel segmentation achieving 97.6% pixel accuracy and 0.82 Dice score on 800 fundus images, ResNet34 deep learning classification with transfer learning on 12,460 ODIR-5K images for binary disease detection, and PCA + Random Forest baseline achieving 99.6% dimensionality reduction across 8 disease categories. Built comprehensive preprocessing pipeline with contrast enhancement, normalization, and data augmentation.

PyTorchResNetU-Netscikit-learnOpenCVPythonDeep Learning
FundusVision: Automated Retinal Disease Detection visualization

Chemical Exposure Risk Assessment App

Mobile application utilizing React Native for cross-platform development, integrated with gaseous diffusion modeling engine and computational algorithms to simulate chemical exposure pathways and calculate risk assessment scores for product safety evaluation. Implemented SQL database architecture and integrated third-party retail APIs to automate safety assessment across 5,000+ retail products with real-time hazard analysis.

React NativeJavaScriptSQLAPI IntegrationMobile Development

Retail Price Database System

Engineered full-stack mobile application using React Native and JavaScript to aggregate real-time grocery pricing data across 15+ retailers. Implemented RESTful API architecture, web scraping, and JSON data parsing to collect pricing, inventory, and geolocation information. Designed SQLite database schema with normalized tables for product cataloging, price history tracking, and user preferences, achieving sub-second performance across 5,000+ SKUs.

React NativeJavaScriptREST APISQLiteWeb Scraping

Multiplayer Video Game

Architected cloud-deployed multiplayer trading card game platform using Django and Python, implementing backend development with game state management and battle mechanics for concurrent user sessions. Collaborated with 4-person team using Agile methodology to deploy web application infrastructure on PythonAnywhere, implementing user authentication, session management, and SQL database operations.

PythonDjangoHTML/CSSSQLCloud Deployment

© 2026 Steven Murley. All rights reserved.