Selected Work
Projects
Deep learning and embedded systems work, mostly at the intersection with healthcare — from hearing aids to clinical audio monitoring.
-
Deep Neural Networks for Speech Enhancement on Resource-Constrained Devices
Designed, trained, and optimized deep learning models in PyTorch and TensorFlow for low-latency noise suppression on embedded devices, targeting real-time deployment on hearing aids.
Read the paper → -
Audio-Based Cough Detection in Clinic Waiting Rooms
A cough detector for clinic waiting rooms built with TensorFlow, identifying cough sounds and estimating the share of patients reporting cough as a symptom — used as a surveillance tool for respiratory disease outbreaks.
Read the paper → -
Personalizing Over-the-Counter Hearing Aids Using Pairwise Comparisons
Developed an optimization technique for hearing aid configurations based on patient feedback and multi-armed bandit algorithms.
Read the paper →