Graduate Student Research Assistant, University of Michigan [resume.pdf, cv.pdf]
My research started in bioinformatics and scientific computing, but much of my work now centers on machine learning systems, especially representation learning, generative modeling, and learned latent spaces for structured data.
I am especially interested in end-to-end research engineering: building datasets and preprocessing pipelines, writing training code from raw PyTorch, and working on video models, diffusion systems, world models, and latent-variable architectures.
Education
University of Michigan 2020 – present
Dept Bioinformatics
PhD candidate
Bowdoin College 2015
BA, Biochemistry with Honors
BA, English Literature
All Publications
Moran, S. P., Feng, F., et al. "Statistically rigorous and computationally efficient chromatin stripe detection with Quagga." Genome Research. 2026. [.pdf]
Sun, C., Zhou, C., Daneshvar, K., ... Moran, S. P., et al. "Conserved long noncoding RNA TILAM promotes liver fibrosis through interaction with PML in HSCs." Hepatology. 2025. [.pdf]
Li, W., Sparks, R. P., Sun, C., ... Moran, S. P., et al. "Screening the human druggable genome identifies ABHD17B as an anti-fibrotic target in hepatic stellate cells." Nature Communications. 2025. [.pdf]
Li, W., Chen, J. Y., Sun, C., ... Moran, S. P., et al. "Nanchangmycin regulates FYN, PTK2, and MAPK1/3 to control the fibrotic activity of human hepatic stellate cells." eLife. 2022. [.pdf]
Daneshvar, K., Ardehali, M. B., Klein, I. A., ... Moran, S. P., et al. "lncRNA DIGIT and BRD3 protein form phase-separated condensates to regulate endoderm differentiation." Cell Reports. 2020. [.pdf]
Selected Recognition
Biomedical Informatics and Data Science (BIDS) Training Award, University of Michigan 2021
National Institutes of Health Intramural Research Training Award (IRTA), NIH 2016-18
Current Technical Interests
Generative modeling, world models, diffusion systems, latent-variable methods, learned state spaces, representation learning, and end-to-end research engineering.
.