At Devster Labs, I contributed to an early-stage Sound Event Detection system using a multi-source audio dataset. I also engineered a snore detection model for a medical screening workflow, where class imbalance guided targeted rebalancing fixes.
For my Accident Detection for Healthcare Management project, I trained and optimized a YOLOv8 model and built an inference pipeline served via FastAPI. The project included MongoDB-backed monitoring and dashboard reporting.
I also built a speech emotion recognition pipeline that reached 93% accuracy across eight emotion classes. As a Teaching Assistant at PAFIAST, I mentor students on Python, machine learning, and data analysis coursework.
