First author. Trained a survival-prediction model on TCGA gene-expression data and tested it on two outside datasets (SCAN-B and GSE96058) to evaluate generalization. Used pathway-level features, Cox regression, binary classification, and SHAP values to surface the biological pathways most predictive of patient outcomes. Appraised by Dr. Eliezer Van Allen (Dana-Farber / Harvard Medical School).

First author. Analyzed clinical predictors and demographic patterns in Long COVID patients using machine learning to identify which factors are most associated with persistent symptoms. Submitted to the National High School Journal of Science (NHSJS); under peer review.

Cross-platform desktop and mobile overlay translating speech, sound, screen, and gesture into accessible output for deaf and blind users. 6 real-time subsystems, sub-second caption latency.

Multi-agent AI system generating complete curricula (units, lessons, assessments) for AP / IB / college / HS levels. 96.4% alignment accuracy.

An AI tool that searches the web, pulls relevant evidence, and organizes it for debate and research use.

Full-stack AI assistant using RAG so users can ask questions and get sourced answers from their own documents.

Published AI recipe app with nutritional analysis, AI-generated food images, and step-by-step cooking instructions.

CNN trained to detect tuberculosis, COVID-19, and pneumonia from CT scans simultaneously with 80% accuracy.

CNN detecting COVID-19 from lung CT scans with 96% accuracy using custom preprocessing and architecture.

Deep Q-learning agent that teaches itself to play Super Mario Bros by trial and error.

State-level placement in the Technology Student Association Software Development category for Relay, a real-time accessibility middleware.

NC State Leadership Conference — Future Business Leaders of America.

Lead programmer for VEX Robotics Team 16885B: autonomous driving, sensor reading, real-time decision making, telemetry.

Placed 4th in the U.S. Congressional App Challenge (NC district) for RecipeSmith, a published iOS app generating AI recipes with nutritional analysis and food imagery.

5–0 preliminary record with 29 average speaker points in Public Forum debate.

Honor society state-level placement.

Awarded at Pleasant Knoll Middle School.

Credential ID: 8757c9af62694d8c9a433c0dbccc20f3