- Developed an application aimed at predicting Alzheimer's and its type using MRI image
- Utilized FastAPI as the central framework to expose prediction APIs, serving as the core intelligence of the application
- PyTorch implementations of ResNeXt and EfficientNet (from scratch) on the Alzheimer MRI Dataset
- Achieved a validation accuracy of 98.90% and training accuracy of 99.10% in 20 epochs using ResNeXt50.
- Developed a training pipeline for developing highly customized solutions to custom tasks with logging.
- Containerized FastAPI application using Docker and deployed on Render.