Reuben Rosen

Used time-series analysis of longitudinal clinical data to analyze heart rate variability (HRV) metrics for early detection and risk stratification of hypoxic-ischemic encephalopathy and related neonatal illness signals. Built clinician-facing tools to visualize patient-specific HRV trends and is developing predictive models for NEC risk using pre-onset HRV features.

Built a web-based server-client application for biomedical image object detection with an easy-to-use GUI for non-programmers to run models, annotate images, and retrain for custom data. Developed a 3-in-one workflow tool that outputs detection coordinates and object type while supporting retraining to improve performance for objects outside the pre-trained model.

Engineered Python-based automation tools for neurological assessment, including computer vision methods for limb detection (pronator drift) and pupillometry automation. Led the team and applied Scrum, Agile, and Kanban practices to improve project efficiency.

Developed a voice-based security identification system using Vector Quantization in MATLAB, achieving 95% accuracy. Extended the work to test breathing-audio COPD detection for an undergraduate project, achieving 97% accuracy.

Developed MATLAB pipelines for infant audio data processing and emotion classification, with results reported as 91% accuracy and accepted for publication. Implemented YOLO-based models for ocular disease detection in eye fundus images, achieving 87% accuracy.

Earned an MS in Biomedical Engineering at Cornell University from August 2023 to May 2024.

Earned a BS in Electrical and Electronics Engineering at Vellore Institute of Technology from July 2018 to September 2022.