---
title: "Undergraduate researcher leverages CARC for molecular modeling breakthrough"
description: "Physics major Sole Samuels used CARC systems to develop radial density encoding, a new molecular representation for machine learning."
type: News
tags:
  - News
  - Research
  - Students
generated:
  by: "claude/fable-5"
  at: "2026-08-29T00:00:00Z"
sources:
  - id: carc-web
    resource: "https://carc.unm.edu/research/molecular-modeling-breakthrough.html"
    title: "Molecular modeling story (carc.unm.edu)"
    author: "team:unm-carc"
---

# Undergraduate researcher leverages CARC for molecular modeling breakthrough

<p class="carc-byline">By Mariah Rosales</p>

For many students, a senior thesis marks the culmination of their
undergraduate studies. For UNM physics major Sole Samuels, it was also an
opportunity to make a meaningful contribution to the field of computational
molecular science — with the help of the University of New Mexico's Center
for Advanced Research Computing (CARC).

Samuels'
[2025 honors thesis](https://physics.unm.edu/undergraduate/thesis/2025_samuels.pdf){target=_blank},
*Radial Density Encoding of Molecular Environments*, introduces a new method
for representing molecular structures that could improve how scientists use
machine learning to predict molecular properties. By focusing on radial
density encoding, Samuels explored how to capture the spatial distribution
of electron density around atoms in a way that is both compact and highly
informative for computational models.

"Training and testing my models required more computing power than I could
ever have accessed on a personal computer," said Samuels. "CARC's
high-performance computing systems allowed me to work with larger datasets,
run more complex simulations, and refine my approach much faster than would
otherwise be possible."

The project bridges physics, chemistry, and data science — demonstrating how
advanced computational techniques can help scientists tackle challenges in
quantum chemistry, such as predicting chemical reactivity or designing new
materials. Using CARC's resources, Samuels was able to train neural networks
on three-dimensional molecular data, a process that demands both significant
processing speed and large-scale storage.

"This experience gave me hands-on exposure to the kind of computing
infrastructure that drives modern scientific research," Samuels said. "It
wasn't just about finishing a thesis — it was about learning how to turn a
theoretical idea into a computational tool."

Samuels' work is a reminder that high-performance computing is not only for
large research teams or faculty-led projects — it is also a critical
resource for students preparing to enter research careers. By providing
access to advanced computing power and technical support, CARC helps equip
the next generation of scientists with the skills and experience needed to
make meaningful contributions in their fields.

![Figure from the thesis](https://carc.unm.edu/research/picture1.png)