Σ
Symmetry-adapted ab initio models
Building SAM, a high-performance framework that exposes U(3) and U(4) content in first-principles wave functions and uses it to compress nuclear many-body spaces.
Research
Discovering, explaining, and representing emergent nuclear symmetries—from first-principles many-body theory to interpretable and equivariant AI.
Atomic nuclei are quantum many-body systems of rapidly growing complexity, yet their low-energy behavior often displays striking simplicity: collective motion, clustering, deformation, and spin-isospin regularities. My research asks how these patterns emerge from realistic interactions and how their underlying U(3) and U(4) symmetries can be used to organize calculations.
I combine representation theory, symmetry-adapted ab initio modeling, high-performance computing, and machine learning. The unifying goal is not simply to predict nuclear observables, but to build methods that reveal why the predictions work—and use that physical structure to guide the next calculation.
Focus areas
Building SAM, a high-performance framework that exposes U(3) and U(4) content in first-principles wave functions and uses it to compress nuclear many-body spaces.
Connecting symmetry-resolved calculations to global nuclear data with WINN, while developing symmetry-aware neural quantum states and exact-to-approximate equivariant architectures.
Developing uncertainty-aware surrogates and agentic workflows that select, run, validate, and interpret informative calculations across the nuclear chart.
Beyond the science, I invest heavily in building scientific communities — founding chapters, leading outreach, and advocating for researchers.
Selected projects
A progression from mathematical infrastructure to first-principles simulation and symmetry-aware AI—each project supplies a layer of the broader discovery program.
An intrinsically interpretable model connecting first-principles symmetry information to global nuclear masses, with accurate predictions, reduced data requirements, and stable extrapolations.
Leading an open LSU AI Journal Club series that develops the mathematical foundations of symmetry and connects them to practical equivariant-learning architectures.
Benchmarking neural representations of nuclear wave functions and extending them toward architectures whose hidden channels and couplings encode exact or approximate physical symmetries.
Main developer of the theoretical framework and HPC C++ software. Current work incorporates three-nucleon interactions through the normal-ordered two-body approximation and extends systematic symmetry-resolved studies toward selected sd-shell nuclei.
Led the LSU team in collaboration with Bard College and MIT for APS's national physics outreach program.
Modernized the Draayer–Akiyama U(3) algorithm and developed the first general U(4) coupling
implementation for nuclear modeling; released as the C++ libraries u3lib and u4lib.
Studied stable nuclear shapes, clustering, and collective behavior via the multiconfigurational dynamical symmetry (undergraduate thesis).
Explored hidden symmetries of the Kepler problem across large exoplanetary databases.
I'm always happy to discuss nuclear structure, symmetry, and AI for science.
Get in touch