Prashant Kumar
PhD Scholar (PMRF) · Aerospace Engineering
Indian Institute of Technology Kanpur
Exploring
Research Summary
My research focuses on Scientific Machine Learning with emphasis on Physics-Informed Neural Networks (PINNs), compressible flows with shocks, multiscale turbulent flow modeling, and GPU-accelerated high-performance computing. I develop data-free and hybrid attention-based frameworks for computational fluid dynamics.
Data-Free PINNs
Solving flow problems from the governing equations alone, without training data.
Shocks & Compressible Flow
Robust physics-informed models for shocks and shock–boundary layer interactions.
Multiscale Turbulence
Attention-based and multiscale networks that capture long-range dependencies.
Turbomachinery
Turbo-PINN, wake prediction in low-pressure turbines, and LES of the SPLEEN cascade.
GPU & HPC
underPINN, a modular JAX-based framework, and GPU-accelerated simulation.
Latest News
Contact
Email: kumarp22@iitk.ac.in
Affiliation: Department of Aerospace Engineering, IIT Kanpur
GitHub: iitkprashant