Nastaran received her BSc from Shiraz University in Iran in Electrical & Computer Engineering in 2013, and after completing her Masters at the University of Oklahoma, joined the UCLA Physics and Biology in Medicine (PBM) graduate program in 2015 where she began conducting her research at CVIB. The research was motivated by the goal of translating radiomics to clinical practice. For this to happen radiomics features must be consistent across different scanners and imaging protocols if machine learning is to be reliable. Nastaran built upon a computational pipeline and body of work by previous students and scientists that allowed her to compute a large number of CT image reconstructions from raw data and measure the variability of radiomics features. Having done this she also explored methods using deep learning, in particular generative adversarial networks (GANs), to transform CT images to conform to a standard acquisition, thereby making derived radiomics features more consistent. The work provides valuable insights and mechanisms for reliable application of radiomics and machine learning in general under the variable real world conditions of imaging in clinical practice.