I Do Science

Who am I

Kaze W. K Wong

Assistant Research Professor

Research Software Engineer

Johns Hopkins University

I am an assistant research professor in the Department of Applied Mathematics and Statistics at Johns Hopkins University. I am also a software engineer with the Data Science and AI Institute. I have very broad interest in many different subjects. In brief, I spend 20-30% of my time thinking about astrophysics, ~40% of my time trying to understand to make neural network robust and how to tune them, and the remaining time building production-grade domain science applications. My work is primarily computational and I care about open source software a lot. See below for some of the topics I am currently working on.

Link tree

Research Interest

Here are a list of my current research interests

GW150914

Gravitational wave astronomy

I am the lead developer or maintainer of a number of gravitational wave astronomy projects, including jim, ripple, JaxNRSur. My research focuses on the development of new algorithms and building open source software to solve data analysis problems in the field of gravitational wave,

flowMC logo

Adaptive MCMC

Markov Chain Monte Carlo (MCMC) is very commonly used in fundamental science to solve data analysis problems. I am the lead developer of flowMC, an open source software package for adaptive MCMC. It is written in jax, has machine learning capabilities, and is designed to solve sampling problems with complicated geometries.

Geometric filters

Understanding Machine Learning with physics

Trained as a physicist, I want to understand the general principles separating a working machine learning model from a non-working one with some physical arguments. This also includes work I am doing to improve the performance of machine learning models with physical principles. This includes infusing machine learning models with symmetries and understanding the accuracy limits of machine learning from a statistical mechanics perspective.

flowMC logo

Machine learning in sport

I am interested in using data science, machine learning, and software engineering to improve the performance of athletes. This includes building computer vision models, extracting analytics from raw data, and create biomechanical models. I also care a lot of about actually deploying the methods and seeing the results as opposed to just publishing.

Research projects

Requirements for working with me

I am trying to be more and more careful in taking in student these days, mainly because since I begun my faculty position, I have received many more requests from students wanting to join my research group. If I say yes to everyone, then it is inevitable that I will spend less time on each student.

I am a person who if I decide to do something, I do it well, and this applies to supervising students. If I consider someone a student of mine, then I will make an effort to make sure the person is happy and improve over time.

Here are some of the criteria that I use to decide whether to accept a student:

  1. I only take students that are either directly recommended by my collaborators or students at Johns Hopkins University. Please do not reach out to me if you do not satisfy this condition. If you violate this condition, I will block your email address, and veto your application whenever I have the chance. Please take this seriously.
  2. Prospective students should have a strong background in computing. This includes but not limited to being proficient in linux, programming languages such as Python, Julia, C/C++ or Rust, and experience with version control. You can still consider joining if you don't have these skills, but you will need to learn them quickly.
  3. Being able to commit at least 10 hours per week and be responsible, including being punctual and reliable.

While I tend to be easygoing in general, I am quite demanding when it comes to work-related matters. I believe professionalism is the most important thing I can teach my students: If I take up a job, I will complete with all of my soul no matter what adversarial conditions I face. Of course I understand things never go as planned and life gets in the way, but I will push you to go beyond your limits and be a better version of yourself. That process is not always pleasant, but I believe this professional mindset will carve a path for you to success.

Open Projects

If you still decide to send me an email to work with me after reading the paragraphs above, we are in game! Here below is a list of open projects I have for new students who are interested in working with me:

  1. Machine learning analysis for sports performance
  2. Development of adaptive monte carlo methods
  3. Observational astronomy data analysis
They are ordered by required skills. Sport analytics being the most beginner friendly and observational astronomy being the most advanced.

Perks you will receive

If you are eligible, and find a project that excites you, then let me further entice you with perks you will receive while working with me:

  1. Access to my GPU cluster
  2. Learn practical computing skills such as MLOps and high performance computing
  3. Opportunity to a publication. I ground all of my research in the context of a publication.