Computer Science Senior at FAST-NUCES Islamabad, interested in machine learning and high performance computing.
arrhythmiaClassifier: final-year project on ECG arrhythmia classification for edge devices. A published LightGBM classifier reaches 99.6% accuracy on MIT-BIH under a random beat-level split, but 76% (anomaly recall 0.30) when tested on unseen patients (inter-patient DS1→DS2 split). Next: compact models trained with knowledge distillation, evaluated only on unseen patients.
zain2466sukhera@gmail.com · LinkedIn
"Uss rizq se maut achi, jiss rizq se aati ho parwaaz mein kotahi." Better to die than live on sustenance that weakens your flight. (Allama Iqbal)