Medical Image Annotation Workflow Participant / AI Engineer
This role involved developing and participating in image annotation workflows for an AI diagnostics platform targeting diabetic retinopathy and glaucoma. I curated and quality-controlled approximately 25,000 OCT volumes and fundus photographs stored as DICOM files, enhancing the dataset for machine learning tasks. Automated annotation throughput was achieved using ITK-Snap and 3D Slicer tools integrated into the data pipeline. • Developed and refined segmentation models for retinal-fluid and anomaly detection. • Used PyTorch and MONAI for image classification and segmentation experiment rounds. • Implemented methods to lift annotation throughput by 40% and ensured dataset quality. • Collaborated with clinical experts to validate annotations and support regulated medical-AI development.