Skip to content
View mikemao27's full-sized avatar
  • Rice University
  • Sammamish, WA
  • 06:19 (UTC -07:00)
  • LinkedIn in/mikemao31415

Block or report mikemao27

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
mikemao27/README.md

Hi, I'm Mike Mao

I use GitHub casually and will occasionally make contributions or commits to projects I find meaningful. I have a wide variety of interests, most of which are a bit "far-fetched" for a brief introduction. I try to produce more quality vs. quantity.


  • 🔭 Double-major in Computer Science & Statistics (B.Sc) at Rice University
  • 🌱 I'm most familiar with Python and Java, however I've made it a goal to expand my tech-stack.
  • 📝 I try to stay updated on LinkedIn, but I am not particularly drawn to "tooting my horn."
  • 💬 Ask me about about some of my interests or check out some of the public repositories I have.
  • ‼️ Please note that not everything I've ever worked on is on GitHub.

Just to keep things direct, my primary interest revolves around AI/ML. Rather than working on the auto-regressive AI paradigm that has permeated industry, I am interested in more philosophically robust forms of AI. This includes, but is not limited to, natural language understanding for English (although the same technology could be applied more easily to inflective languages like Latin: which use explicit word endings to communicate gender, number, and case, as well as relationships between words), or spiking neural networks (SNN). Many of the projects I work on are run-able on simple hardware like a MacBook Air (M4 Chip) and should not require extensive GPU capabilities. I try to make the cost requirements for my work as low as possible (not just for others, but also due to my self-imposed financial restrictions), meaning that much of these projects required $0 (or near $0) to develop: often yielding what are essentially 0 parameter models (yes, this is possible if you abandon certain paradigms of auto-regressive black-boxes).

Important

It is my personal belief that the point of technological innovation is not to accumulate vast quantities of capital or fame. Those may naturally follow major technological innovations, but I do not personally feel as if they are the reason why we should innovate. A major part of innovation requires collaborations between groups and ideas that may not have otherwise seemed possible, as well as transparency and access. Thus, I view the open-source movement to be especially important to the unending development of equitable technology. As such, I make most, if not all, of the projects I personally feel to have a met a certain bar of quality open-source and public for all to use (with a license attached: often-times an Apache 2.0 license).


Contribution Graph

Pinned Loading

  1. RouteWorks/RouterArena RouteWorks/RouterArena Public

    RouterArena: An open framework for evaluating LLM routers with standardized datasets, metrics, an automated framework, and a live leaderboard.

    Python 118 34

  2. DEMES DEMES Public

    Deconstructive Encoding of Meaning and Expressive Syntax (DEMES) for the English natural language. Model understanding is accomplished by encoding English primitives mathematically in a world model…

    Python 1

  3. KINETIC KINETIC Public

    KV-Informed Neural Inference and Token-Skipping Execution Core (KINETIC) for diffusion-based large-language model wall-clock speedups. The baseline model used is NVIDIA's Fast-dLLM v2.

    JSON 1