Back to Portfolio
Research Project
2023 Completed
AI-Powered ECG Signal Classifier
Deep learning model that classifies arrhythmias from short-window ECG with high precision suitable for edge devices.
- Role
- Researcher
- Client / Partner
- Collaborative Research with Hospital Partner
- Year
- 2023
Problem
Manual ECG interpretation is time-intensive and limited by specialist availability.
Solution
1D-CNN with attention pooling deployed via TensorFlow Lite for low-latency inference.
Output
Journal publication, deployable model artifact.
Tech Stack
Python
TensorFlow
NumPy
TFLite
Edge AI