At TCG Crest, I compared deep learning architectures for medical image segmentation in head and neck oncology. On a funded research project, I implemented and evaluated Attention U-Net, U-Net++, U-Net+++, and nnU-Net on an HNSCC MRI dataset using PyTorch.
I also worked with pre-clinical mouse brain MRI and clinical T2-weighted head and neck oncology datasets, preprocessing 3D volumes into 2D slices for GPU efficiency. I assessed segmentation accuracy by freezing encoder layers and retraining decoders, and assisted with manual cell segmentation and image orientation correction for zebrafish images.
My projects include a content-based movie recommendation system and a Python recommendation engine for the e-commerce website Apparalls. I also analysed employee attrition as a Data Analyst Intern at Unified Mentor and co-authored a systematic review on AI and ML in cancer treatment.
