Nisha Chandramoorthy

Timeline

1. 2014–2016

   IIT Roorkee and an MIT master's

   She earns a bachelor's degree in mechanical engineering at the Indian Institute of Technology, Roorkee, with a 2014 thesis on the fast multipole method in particle vortex methods. In 2016, she completes an MIT master's thesis on molecular dynamics-based approaches for mesoscale lubrication, supervised by Nicolas Hadjiconstantinou.

2. 2019

   AIAA Aviation best student paper

   She wins the best student paper award at AIAA Aviation 2019 for "Sensitivity computation of statistically stationary quantities in turbulent flows," written with Qiqi Wang.

3. 2021

   A PhD on chaotic sensitivity

   She completes her MIT PhD in 2021 at the Aerospace Computational Design Lab, advised by Qiqi Wang. Her thesis, "An efficient algorithm for sensitivity analysis of chaotic systems," proposes space-split sensitivity (S3), which evaluates the linear response of a chaotic system as an efficiently computable ergodic average.

4. 2021–2022

   Postdoc at MIT IDSS

   From fall 2021, she is a postdoctoral researcher at MIT's Institute for Data, Systems and Society, mentored by Youssef Marzouk and Stefanie Jegelka. She works on improving filtering of chaotic dynamical systems and on the dynamics of learning algorithms. In 2022, she publishes a linear-response paper with Qiqi Wang in the SIAM Journal on Applied Dynamical Systems and a justification of the space-split sensitivity algorithm with Malo Jézéquel in Nonlinearity.

5. 2022

   Statistical algorithmic stability at NeurIPS

   Her paper "On the generalization of learning algorithms that do not converge," written with Andreas Loukas, Khashayar Gatmiry and Stefanie Jegelka, appears at NeurIPS 2022. It proposes statistical algorithmic stability (SAS), which extends classical algorithmic stability to non-convergent training, and gives evidence that networks that train stably generalize better.

6. January 2023–June 2024

   Edenfield professor at Georgia Tech

   She works as a James C. Edenfield Early-Career Assistant Professor in Georgia Tech's School of Computational Science and Engineering from January 2023 to June 2024. In February 2024, she gives a joint University of Chicago colloquium for the Department of Statistics and the Committee on Computational and Applied Mathematics, titled "Learning from dynamics, learning dynamics and the dynamics of learning."

7. August 2025

   Inexact generative models paper

   In August 2025, she and Adriaan de Clercq post "When and how can inexact generative models still sample from the data manifold?" The paper takes a dynamical systems approach and shows that the alignment of the top Lyapunov vectors with the tangent spaces along the boundary of the data manifold leads to robustness of the support.

8. As of October 2026

   Assistant professor at UChicago

   She is an assistant professor in the University of Chicago's Department of Statistics and its Committee on Computational and Applied Mathematics. One focus of her research is sampling algorithms for Bayesian posteriors that accelerate inference within dynamical systems, and her group is working on the mathematics of generative modeling and sampling.