Latent evolution models
Spatiotemporal learning of charged-particle beam dynamics
Forward and inverse modeling, uncertainty quantification, and tuning of charged-particle beam dynamics in particle accelerators.
Read the Physical Review E paper →Scientific machine learning · Los Alamos National Laboratory
Mahindra S. Rautela
Postdoctoral Researcher
Teaching physics to machines.
I develop scientific machine learning methods to solve forward and inverse problems in high-dimensional spatiotemporal dynamical systems.
Background
About
I am a Postdoctoral Researcher at Los Alamos National Laboratory. I earned my Ph.D. from the Indian Institute of Science, Bangalore, in June 2023, with a significant portion of my research conducted at Purdue University as a visiting student.
My research focuses on scientific machine learning for forward, inverse, and optimization problems in spatiotemporal dynamical systems. Current work includes PDE foundation models, operator learning, and latent evolution models for particle accelerator dynamics.
Positions
Postdoctoral Researcher
Los Alamos National Laboratory
2023–present
Project Engineer (Research)
IIT Kanpur
2018
Education
Visiting Ph.D., Mechanical Engineering
Purdue University
Jan. 2022–June 2023
Ph.D., Aerospace Engineering
IISc Bangalore
2019–2023
M.S./M.Tech., Aerospace Engineering
IIST Trivandrum
2016–2018
Current work
Research
Physical systems
PDE foundation models
Foundation models that learn across data modalities and transfer to physical systems outside the training distribution.
Read the MORPH paper →Inverse problems
Physics–data models
Machine-learning surrogate solvers and optimization methods for mechanics-based inverse problems.
Explore all research projects →Latest
News
Selected as a Gold Reviewer for the ICML 2026 Conference.
Out-of-distribution transfer of PDE foundation models to material dynamics under extreme loading was accepted to the ICLR 2026 AI & PDE Workshop. Workshop →
PDE foundation model-accelerated inverse estimation of system parameters in inertial confinement fusion was accepted to the HPAI4S workshop at IPDPS 2026. Workshop →
MORPH: PDE foundation models with arbitrary data modality was released on arXiv. MORPH paper →
Research output
Recent publications
Complete archive (25 articles) →Under review
CBOL-Tuner: Classifier-pruned Bayesian optimization to explore temporally structured latent spaces for particle accelerator tuning
Community
Service & recognition
Reviewing
Reviewer for more than 30 articles across journals including Nature Machine Intelligence, Scientific Reports, and IEEE, IoP, Elsevier, ASME, and Springer publications.
Gold Reviewer, ICML 2026; reviewer for ICLR 2026 workshops.
Recognition
Prof. Chintakindi V. Joga Rao Medal for best Ph.D. thesis, Indian Institute of Science, 2024.
Overseas Visiting Doctoral Fellowship, SERB-DST, 2022–2023.
Talks
Invited talks on generative modeling, particle accelerator dynamics, and data-driven plasma science.
View selected talks →