Announcing a series of blog posts spanning machine learning, climate science, galactic winds, and cosmic rays!
You’re probably thinking, “None of these topics are connected, and some of these words sound made up…this blog is going to be an incoherent mess, and the author must be too!”
Well, let me explain…
From Galaxy Atmospheres…
I’m a computational astrophysicist by training. I got my PhD in Physics in 2020 from University of Wisconsin-Madison simulating the gas flows in and around galaxies. I particularly focused on how galaxies can drive large-scale outflows (“galactic winds”) into their surrounding atmospheres (the circumgalactic medium). For 6 years, I designed and built new astrophysics modules and integrated them with massively parallel C++ and Fortran-based codebases, ran large-scale models on some of the country’s largest supercomputers, and analyzed terabytes of data using Python-based data science tools (e.g. Numpy, Pandas, Scipy).
I then did a 3-year postdoctoral fellowship at the Kavli Institute for Theoretical Physics at UC-Santa Barbara, where I was the principal investigator (PI) on a supercomputing grant (> 15 million CPU hours) supporting multiple postdoc and PhD projects. Our main objectives were to better understand how cold clouds in galaxy atmospheres survive fluid instabilities and “rain” down on galaxies, and how relativistic charged particles called cosmic rays can significantly alter galaxies and their surroundings despite comprising only a billionth of all particles in the Universe.
Most recently, I’ve created a >10 terabyte simulation suite exploring the interplay between cosmic rays and magnetized turbulence (Bustard and Oh 2022, 2023), and using those images as training data, I’ve developed a novel technique to reveal cosmic ray properties from gas density images using convolutional neural networks (Bustard and Wu, in prep). This project is ready for journal submission, so look for a blog post about this very soon!
…To Lots and Lots of Self-Reflection…
I’ve greatly enjoyed my time as a computational astrophysicist, especially the diverse and intellectually stimulating collaborations and the opportunities to sculpt and advise projects for 6 junior researchers; however, I’ve also become more aware of my personal priorities that are at odds with the traditional academic route, e.g. having flexibility in work location (ask me about New Mexico and Wisconsin, two under-appreciated states I plan to spend a lot of time in), creating positive contributions to society beyond the accumulation and dissemination of knowledge, and maybe contributing a bit more financially to our family so my partner doesn’t have to bear the entire load any more.
For those reasons and others, I recently decided to forego a career in academia and test the job market for data science, machine learning, and software engineering roles! Instead of all roads leading to tenure-track, I’m now exploring dozens of career paths, which on one hand is very overwhelming, but on the other hand, very freeing and exciting!
…To Saving Earth’s Atmosphere
While I’m genuinely enthused to tackle complex, ambiguous problems of all sorts, one path in particular stands out: climate science, particularly data-intensive solutions to climate change such as optimized transportation, smart use of the power grid, and machine learning-informed verification and monitoring of corporate greenhouse gas emissions. During the current explosion of AI in our daily lives, from chatbots to pointed advertisements, I’m optimistic that AI can also support our necessary fight to offset global warming.
To that end, I recently participated in a series of lectures and tutorials at the Climate Change AI 2023 Summer School, where I greatly expanded my knowledge of clean energy funding and regulations, cutting-edge climate science research, and use-cases of machine learning in this space, from natural language processing to remote sensing and computer vision.
Starting September 25th, I’ll be one of the few hundred Climatebase Fellows (Cohort 4) accepted out of a pool of a few thousand! I’ll spend 12 weeks connecting with other like-minded individuals from diverse academic and professional backgrounds, learning about global warming and efforts in the tech industry to slow and reverse the warming trend, and devising and executing a collaborative capstone project to present to industry professionals later this year!
Why create a blog?
I’m naming this blog “From Galaxy Atmospheres to Saving Earth’s Atmosphere” partly because it’s catchier than “My Blog” but mainly because I hope it captures a story of transition from theoretical astrophysics research to applied climate technology, an area I imagine many other STEM and non-STEM PhDs would be thrilled to join but probably don’t know where to start. As I’ve found out over the last 2 years, making a career transition is extremely difficult. It first requires you to divest from your one-track career path and your past identity, which may have been engrained in you for many years; this part is the hardest and takes the longest. It then requires you to embrace uncertainty and ask others for help; this part is uncomfortable. Then it requires you to pour over your previous career and remind yourself what you’ve done, put all of that in a presentable resume, cover letter, blog, etc. and start applying; as of writing, this is my current phase.
I hope that 5 years from now, when I’m an established climate technologist leading a successful group of diverse, remote, and enthusiastic data scientists, a data-oriented PhD will find this blog, feel empowered to make a scary career switch, learn a thing or two about the transition to industry, and see the exciting steps we’ve taken towards off-setting global warming. For now, this blog is meant to show a bit of who I am, a bit of what I’ve done as an astrophysics researcher, and a bit of what I’m currently learning as I explore the realm of clean energy and AI-informed solutions to global warming. Cheers to you for reading this far!
Above is a beautiful stock image (which looks surprisingly similar to my current desk setup) for a Quarto-generated blog post. Quarto allows you to convert things like Jupyter notebooks to pretty blog posts hosted on a variety of platforms, such as GitHub Pages.