About Me

I am a postdoctoral researcher and a member of the Solar System Dynamics (SSD) group and the Center for Near Earth Object Studies (CNEOS) at the NASA Jet Propulsion Laboratory. I work on data pipelines and software for SSD and CNEOS, helping to provide robust asteroid ephemeris and impact monitoring services to the public and the astronomy community in the era of next-generation survey telescopes, including the Near-Earth Object Surveyor and the Legacy Survey of Space and Time.

As a graduate student, I studied astronomy at the University of Washington. I worked on various projects related to data science and applications of scalable computing in astronomy, with a focus on preparing for the Rubin Observatory's Legacy Survey of Space and Time with my advisor Prof. Mario Juric and collaborators. I am an alumus of the Computational Science Graduate Fellowship program, which supported my studies from 2018 - 2022. Through the fellowship program, I worked with Dr. Michael Grosskopf, Scientist at Los Alamos National Laboratory, on emulating nuclear physics simulations using scalable modeling techniques. My PhD thesis, entitled "Driving Discovery in Astronomy Using Scalable Computing and Fast Algorithms" can be accessed via this website and the UW library sytsem.

As an undergraduate physics and computer science major at the University of Virginia, I worked in high energy physics with the Mu2e collaboration, mentored by Prof. Craig Dukes and Prof. Craig Group. During an internship at the National Radio Astronomy Observatory, I studied binary pulsars with Dr. Kevin Stovall and with his help formed the Pulsar Observers Research Group at Virginia.

Feel free to reach out to me via email if you want to get into contact. Also find a copy of my CV at the link below and an ADS link.

Email    CV    ADS    GitHub

Research: Cloud-Based Science Platforms

With my advisor Mario Juric and the Data Management group at UW, I developed software for deploying "science platforms" in cloud computing centers. These science platforms are based on Jupyter Notebook servers spawned from a JupyterHub, each running as Docker containers on a Kubernetes cluster. Integrated within the Jupyter Notebooks are tools based on Apache Spark for analyzing TB+ sized datasets. We've deployed several of these platforms to:

  1. support analysis of catalog data from the ZTF,
  2. to support the 2020 GROWTH summer school,
  3. to support the 2020 Astro Hack Week,
  4. and most recently three science collaborations for the LSST.
Find at the following link more detail about this work including publications and talks.

Publications and Code

To further develop the usability of these science platforms, I worked with Mario Juric to develop a JupyterHub that utilizes the Checkpoint/Restore functionality of Podman containers for transparent migration of Jupyter notebooks between servers. This allows the user to access more computing resources (CPU and Memory) without having to restart running code.

Paper    Code

Finally, I mentored Biswarup Banerjee for the 2020 Google Summer of Code (GSoC), developing an extension to the Jupyter notebook interface to allow for easy creation and manipulation of Apach Spark clusters, which we call the "sparkmanager."

Code

Research: Binary Pulsars

With Dr. Kevin Stovall at the National Radio Astronomy Observatory, I analyzed the NANOGrav 11 year data set to infer population level statistics about binary pulsars. Specifically, we investigated the claim that binary pulsars are distributed on the sky uniformly over the cosine of their inclination. Research products produced through this work include a poster for AAS meeting 231, and a summary document for internal NRAO use.

Poster Abstract    Poster PDF    Report PDF

Image courtesy of NRAO/AUI.

Project: Performance Analysis of Approximate Gaussian Process Regression

For the University of Washington course CSE 547 "Machine Learning for Big Data" I produced a performance analysis of Kernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP) in the context of kernel (or hyperparameter) learning. I use this method to infer the rotational period of a variable star, following the procedure of Foreman-Mackey et al. (2017). Below you can find a poster and summary document for this project.

Summary PDF    Poster PDF

Project: The Musical Wayfinder

For the University of Washington course CSE 512 "Data Visualization" I worked with Hannah Bish to design, create, and publish an interactive tool for exploring Spotify music libraries using the d3.js JavaScript data visualization library. Below you can find a link to the code used and the project page which includes links to a summary poster and paper. The project page also lets you visualize your own library.

Project Page    Code