Senior Software Engineer
Current* Developed 3D z-stacked, independent channel YOLO based deep learning algorithm for fluorescence in situ hybridization (FISH) signal count differentiation, with significant ablation studies showing it was significantly superior to more generic methods. Also advised data labeling and validation efforts in the project, which was tricky due to the high level of noise in the data (debris creating false signals, faint signals due to various issues, etc). * Designed and built an initial diverse dataset for training a novel cytology foundation model, that was small enough to develop in fast cycles and big enough to use powerful models on.