Hi, I'm Daniel!

I'm interested in understanding how AIs think.

To that end, I currently work at a company I founded, applying AI, and at OSU NLP, researching AI.

My applied work involves:

  1. Building systems to collect lots of data.
  2. Building agents for long-horizon and complex tasks, such as research support for basic science.

My research interests relate to:

  1. Recurrent depth.
  2. Understanding how AIs build internal representations of agency.

My hope is that this research will allow us to build AIs that are more capable, and more aligned.

Previously, I worked on applied AI for healthcare, was a Millennium Fellow, and briefly explored TQFTs. I recently received my undergraduate degree from the Ohio State University, where I studied mathematics and the anthropology of science and technology.

I'm on the organizing committee of Cycle, OSU's undergraduate math research program, and I sit on the Advisory Board of the OSU Department of Mathematics.

If you're interested in what I'm working on, feel free to reach out!

paper

Improving Efficient Recurrent Depth Via Next-Latent Alignment (opens in a new tab)

Workshop poster, with Yuekun Yao and Huan Sun

First page of Improving Efficient Recurrent Depth Via Next-Latent Alignment