Resume

A PDF version can be found here .

Kit Lee

web: wingkitlee0.github.io
email: wklee4993(at)gmail.com

Education

University of California, San Diego
PhD in Physics (Astrophysics)
Oct 2013
Thesis: Spiral Shock and Feathering Instabilities in Spiral Arms

Chinese University of Hong Kong
BSc in Physics
Sep 2007

Technical Experience

Abstract Analyzer

2018 – present

  • A machine-learning web-app that can classify user’s summary of a paper into one of the astronomy categories. The classification task is done using the tensorflow library hosted on the serverless Amazon Lambda platform.
  • Technologies: tensorflow, NLP

Alexa Skills on Amazon Echo devices (astro-ph)

2017 – present

Open Source Contributions

2018 – present

scikits.bvp_solver

  • A Python library for solving two-point boundary value problems, based on a well-known Fortran library.
  • Technologies: Fortran, f2py, setuptools

scikit-cuda

  • A Python wrapper for the MAGMA GPU linear algebra library, which allows drop-in replacement of scipy.
  • Technologies: cuda, cusolver, blas, lapack

Arxiv Reader (Mobile)

2016

  • An app to browse papers via the RSS feeds from arxiv.org. Over a hundred of acquisitions.
  • Technologies: .NET, C#, Visual Studio

Freelance Projects

2009

  • A large Fortran application was parallelized using OpenMP. Resulted in nearly linear-scalability of the program on multi-core machines. Received a 5-star review for the project.
  • Technologies: Multi-threaded debugging, Fortran

Academic Experience

Northwestern University, Evanston, IL

Postdoc | Aug 2016 – Aug 2019

  • Perform research on fluid dynamics of celestial objects, such as forming-planets and galaxies.
  • Develop object-oriented C++ code to solve fluid problems based on the finite-difference method.
  • Present published results in conferences/seminars as invited speaker.
  • Awarded Microsoft Azure Research Grant in 2017 for code development ($20000)

Academic Sinica Institute of Astronomy and Astrophysics, Taipei, Taiwan

Postdoc | Oct 2013 – Aug 2016

  • Developed a Python program to perform parallel execution of Fortran code on the multi-node environment.
  • Participated in team development of a high-performance computational fluid dynamics code. Developed a parallel solver for Poisson equation using Fast Fourier Transform.
  • Awarded a postdoctoral fellowship and a national paper award in 2015 for the project.
  • Organized an international workshop and a hack day.
  • Taught a course on Python in a summer school.

UC San Diego

Research Assistant | Aug 2008 – Sep 2013

  • Created a new numerical algorithm to solve the coupled nonlinear differential equations.
  • Awarded 200,000 cpu-hours from national computing facilities (Teragrid) to perform high-throughput parameter-sweeping calculations.

Additional Information

Languages: Chinese (Native), English (Professional)
Technical Skills (selected):

  • C++, Python, Git, bash shell, Fortran, Linux, C, CUDA
  • Experience in HPC (OpenMP, OpenMPI), cloud computing (AWS, Azure, GCP), and machine learning (tensorflow, keras, scikit-learn)
  • Others: Mathematica, docker