About Me
A Bit About Me
I’m a small-town earthy-crunchy hacker, passionate about coding, machine learning, teaching, and tinkering. For me, programming is more than just a career: Programming is my art, and computers are my canvas. Computers are a space for play, an endless area for exploration, and a medium to sculpt to my personal tastes. If there is one thing I have found I love most in life, it is learning and the constant pursuit of curiosity and knowledge.
Nearly everything I’ve tinkered with, hacked on, coded, and created has been a labor of love, passion, and a deep expression of my endless curiosity. At this point I believe it’s safe to say that programming has been, and hopefully will remain, my lifelong passion.
Aside: My “AI” Use Policy
I take great pride in my craftsmanship, with learning through my work as a cornerstone of that. As such, I do not use LLMs or “AI” to ghost-write my work or assist in any meaningful way. I extremely rarely, if ever, use LLMs when coding, writing, or researching, and when I do the use is very limited in scope. I have many moral, technical, privacy-related, and artistic qualms with the use of LLMs’ in society. Even if I overlook my moral objections and hesitancy based on my ML Theory Research (e.g., adversarial attacks, interpretability), I think using these tools would cheapen my craft and rob me of many opportunities to learn.
All that said, I know that LLMs role in society cannot be overlooked, and I firmly believe that it’s responsible use is an important topic to research and disseminate. I have written a bit about this in the “The Trade Offs Users Face” section on my “Fuzzy Finding” blog post already, but I fully intend to publish dedicated, non-technical material on this topic.
Short Bio
I began making my mark on the internet around 2010 when I began writing apps in Visual Basic, writing programs for jailbroken iPhoneOS, and writing blogging software in PHP. Over the past several years, and my thousands of hours making and breaking many forms of software, my deeply curious nature drove me to play in many different domains of computing which have contributed to the causes I support.
I followed my passion for programming into my Bachelor’s, where I pursued a double major in Honors Mathematics and Computer Science. In 2017, I began my journey into the world of Machine Learning Theory Research, where I now continue my investigations into the mathematical machinery that explains the behavior of modern Deep Learning models. For more information on some of my research interests see the Research Interests section below.
Since graduating, my career has been split in two directions: one building upon my skills in Machine Learning research & implementation, and the other leveraging my experience in software engineering. Regardless of what my current “job title” may say, my passion for research and creation continues outside of my job.
Causes I Am Passionate About
I am someone who can easily become passionate about a topic or cause, so pinning down a specific set of causes & topics I am passionate about can be hard. You may find me rambling about my current set of passions to friends, occasionally on social media (albeit rarely used), and maybe in a blog post of mine (if I ever publish some of my many 95%-done draft posts).
The most notable causes on my mind are (in a non-exhaustive and unordered list):
- Sustainable computing
- Permacomputing, and designing efficient and durable software.
- Sustainable open-source software (perhaps we should think more about “free as in freedom” as opposed to “as in beer”…)
- My “Anti-AI AI Researcher Club” mentality – pondering empowering, efficient, and accurate architecture – that was born from my experience in ML Theory Research (also see my Research Interests below)
- Reclaiming the “personal” aspects of personal computing, and studying the early design decisions in designing computers to be accessible and understandable to a general audience.
- Personal computing came from a place of trying to make computers empowering and playful. How can we (as implementers, users, and computer fans) reclaim the spirit that Smalltalk and the early web brought with them?
- Fostering the ideal of an accessible, people-focused, “indie” web
- Sustainable living outside of computing (perhaps some form of Solarpunk?)
- Pubic Libraries & Libraries of Things
- Support community aid & services
- Universal and accessible knowledge, education, and research
- I love to teach and tutor too! Hopefully I can make time for more of that in the future.
Research Interests
Since leaving my undergraduate university and working as a Machine Learning Scientist, my research interests in Applied Mathematics and Machine Learning have crystallized over time from my more naive days as an undergraduate researcher.
I am broadly interested in the intersection of Applied Mathematics and Computer Science, and its applications across diverse disciplines and industries. In particular, most of my current and former research has revolved around Deep Learning Theory and it’s applications to the sciences, although that focus is largely due to my narrow research and industry experience not matching the breadth of my interests.
I hope that one day my research leads to a paradigm where Deep Learning is as interpretable as code and equations, accurate enough to do justice to the sciences and logic, and efficient enough to be democratic in its deployment.
Below are some of the topics that I am interested in, but there are many others I can speak on.
Topics in ML Theory
The research that I find most compelling in machine learning and AI theory is that which is mathematically rigorous, dataset independent and heavy utilizes algebraic constructions and analysis. I can be convinced of properties like “intelligence” in a system if you can prove to me in the math that this behavior emerges, otherwise I view Deep Learning as a mathematical tool.
Some specific topics I am interested in are:
- Aligned & Equivariant ML (e.g. how can we align our models with our scientific knowledge?)
- Efficient AI/ML (e.g. choosing architectural/training primitives that are compute and data efficient)
- Constructive AI/ML (e.g. how can we intentionally construct models for specific behavior?)
- Sparse Autoencoders and Neural Information Theory
- Neural Collapse, Grokking, and Neural Circuits
- Chain of Thought, Neural Computations, and Formal Languages
- Interpretable and “provably intelligent” AI
Broader Research Interests
- Kolmogorov complexity (my favorite topic!)
- Algorithmic and non-algorithmic information theory
- Graph theory, random walks, Markov chains
- Among other things, I would really love to learn more theoretical Physics!!
Random Things I’ve Worked On Recently
The list of projects I’ve recently been working on constantly changes, but here are some things I’ve been toying with recently:
- Building a robust portfolio for future PhD program applications, including building an academic social network and joining research groups as an external contributor.
- Improving my personal knowledge and task management system, that I have built in Emacs.
- Consuming numerous books and research papers daily. Some literature on personal computing, some on ML Theory, and some on programming language design.
- Pondering and laying the groundwork for various improvements to my website structure – such as a “public drafts” folder for blog posts, memex-like hyperlinking for topics that I mention, and a more automated system for keeping my Bookself page more up to date with what I’m actually reading.
- Building apps for e-ink devices. (More information coming soon, perhaps!)
- Finally (!!) finishing some of the draft blog posts I have in a queue.
- Exploring the realm of building apps that run over SSH.
Want to chat or collaborate? Feel free to send me a message, follow me on social media, or put some time on my calendar to meet over on my “Contact Me” page.