Snigdaa Sethuram

During her time at the Center for Computational Astrophysics, Snigdaa S. Sethuram worked on creating a machine learning (ML) algorithm to predict spectral energy distributions (SEDs) of simulated galaxies given galaxy properties, rather than having to run radiative transfer (RT) code on simulations to obtain the SED. Getting SEDs from simulations makes it possible to directly compare simulated objects to observed objects. Running RT calculations is computationally taxing and infeasible for large simulation suites, so the goal of the project was to reduce this computational cost by obtaining the resultant SED using an ML algorithm instead of RT code. She worked with Flatiron Institute scientists Chris Hayward and Rachel Cochrane.

Sethuram received her bachelor’s degree in astrophysics from Rutgers University and her master’s degree in physics from Georgia Tech. She is currently a Ph.D. candidate in the physics program at Georgia Tech and works in John Wise’s Computational Cosmology group. She is also a FINESST fellow.

Since leaving the Flatiron Institute, Sethuram has returned to Georgia Tech to finish her Ph.D. As a graduate researcher, she is focusing on computational cosmology and machine learning, specifically using machine learning algorithms to emulate subgrid star formation and feedback (SFF) calculations and super-resolve cosmological simulations. The goal is to relax the resolution of large-volume simulations and use an ML algorithm to emulate subgrid SFF physics, and then super-resolve the simulations in post-processing as necessary.

She is also a graduate mentor for first-year students. She served as vice president, from 2020–2021, and president from 2021–2022, of the Graduate Association of Physicists at Georgia Tech. She was also a graduate representative for the Georgia Tech School of Physics Graduate Committee from 2021–2022, a member of the Dean advisory committee from 2022–2023, and a Graduate Judiciary Council member from 2021–2022.