Machine Learning Engineer- ORISE Fellow
Food And Drug Administration (FDA)
Aug 2023 - May 2025 (1 year 10 months)
Designed and implemented automated radiology report evaluation methods for LLMs to improve consistency of assessments. Explored if an LLM can serve as an alternative to human experts in evaluating LLM-generated X-ray report. Conference paper (currently under review in FDA). Physics-Informed Data Augmentation (PIDA) to simulate low dose CT scans. Designed and developed a deep learning based PIDA model to synthesize low-dose noise characteristics from high-dose CT scans. Paper under peer review in Medical Physics journal. Github Code