Education

• Ph.D in Civil and Environmental Engineering, Duke University, 2020
  • My dissertation is broadly focused on developing spatiotemporal Bayesian models integrating strong physical priors with flexible parametric specifications. From an application standpoint, my main interest is in understanding and monitoring large-scale changes in the abundance and dynamics of inland water bodies such as wetlands and ponds across North America. I leverage methods and ideas from Bayesian machine learning as well as engineering science to help provide both accurate and reliable forecasts in problems with sparsely observed and noisy data.
• M.S. in Statistics, Duke University, 2020
  • As a dual-degree student in engineering and statistics, I have had the opportunity to conduct research in applied statistical methodologies. I have developed novel combinations of low-rank matrix factorization and generalized linear models appropriate for very high dimensional \((D>10^3)\) spatial non-Gaussian multivariate data with GPU-accelerated inference. I also am interested in developing rigorous and principled methods for uncertainty quantification using deep generative models such as variational autoencoders and generative adversarial networks in conjunction with Markov chain Monte Carlo and variational Bayes.
• B.A. in Physics, Macalester College, 2013

Work & Research

• Fall 2014 - current: Graduate Research Assistant
• Fall 2019: Instructor
  • Model-Based Data Science (Duke CEE690)
• Summer 2019: Research Intern at IIASA
• Fall 2016: Graduate Teaching Assistant
  • Probabilistic Machine Learning (Duke CS571 / STA561)
• August 2013 - May 2014: Database Engineer at Epic
• May - August 2012: Research Assistant in condensed matter physics in Strzhemechny group, TCU

Grants, Fellowships, and Awards

• IIASA Young Scientist Summer Program Fellowship
• NASA Earth and Space Science Fellow (2016 - present)
• James B. Duke Fellow (2014 - present)
• NSF IGERT Trainee (Duke WISeNet, 2014 - 2017)
• Duke Wetland Center Grantee (2016 - 2017)
• NVidia GPU Grant (2017)
• Nicholas School NPAC Grant (2017)
• Amazon Web Services Cloud Research Grant (2016)
• NSF Research Experience for Undergraduates fellow (2012)
• National Merit Scholar (2009 - 2013)

Publications

1. Identifying wetland consolidation using remote sensing in the North Dakota Prairie Pothole Region (C.K., Mark Borsuk and Mukesh Kumar; Water Resources Research 2018)
2. Probabilistic programming: a review for environmental modellers (C.K. and Mark Borsuk, Environmental Modelling and Software 2019)
3. Gradient based inverse estimation for a rainfall-runo ff model using a deep learning optimization framework (C.K., Mark Borsuk and Mukesh Kumar, Water Resources Research 2019)
4. Crop yield response to water availability in the U.S. over the past thirty years (C.K., Emily Burchfield, Danielle Touma, Max Stiefel, Rui Zhu and John Nay, under review)
5. A spatial community regression approach to exploratory analysis of ecological data (C.K., Mark Borsuk, Methods in Ecology and Evolution 2020)
6. A parsimonious Bayesian representation of non-floodplain wetlands (C.K., Mark Borsuk, Mukesh Kumar, under review)
7. Quantifying uncertainty in remote sensing data with deep generative models (C.K., Mark Borsuk, Mukesh Kumar, in preparation)