Erik Peterson

I'm a scientist.

I’m an engineer. I’m a very interdisciplinary mathematician. I’m a Principal Scientist at phinyx.ai.

At phinyx, I lead research and development for a natural language program synthesis product for scientific computing and machine learning.

At Pastuer Labs I led development of surrogate models for multi-physics, automated causal discovery, and other problems at the intersection of physical science with AI, aka scientific machine learning or “simulation intelligence”.

I’m also interested in physical computation, astrocyte computation, single cell learning, open-endedness, and other kinds of unconventional things.

In the past I was a research scientist at CMU where I studied exploration, play, and curiosity and reinforcement learning.

I worked at Kernel.co as a research scientist where we blended biophysical modeling with deep neural networks.

I was a postdoc at UCSD where I worked with Brad Voytek. We used simulations to better understand costs and benefits of neural oscillations. Some neural data science too.

I did my PhD with Carol Seger. We used fMRI, machine learning, and reinforcement learning to study human behavior.

Farther back I've worked in nanotech, chemistry, and biochemistry, including a couple years in the Bay Area Biotech industry.