I am a Ph.D. candidate at Johns Hopkins University working in the Flow Physics and Computation Laboratory with Prof. Rajat Mittal.

At Hopkins, my research is a unique blend of applied mathematics, computer science, and computational mechanics. I primarily focus on developing mathematical and numerical models using differential equations and linear algebra that describe the multiphysics phenomena in nature. These models are extremely large and computationally expensive to solve. The solutions are often obtained by developing sophisticated computer codes using data structures and algorithm concepts. This requires significant programming, debugging, and optimization in C++, Fortran, and MPI to run them extremely fast on HPC clusters utilizing thousands of CPUs to produce simulation results.

Once the simulation is complete, the next task is to harness the chaotic terabyte-scale time-series datasets with at least 200+ million spatial data points, each with at least 4 degrees of freedom. Analyzing this data and extracting meaningful engineering insights is a highly complex task, and that's where I put my statistics and AI/ML skills using Python libraries like NumPy, SciPy, TensorFlow, PyTorch, PySpark, and Pandas to work.

These codes are versatile and can mathematically model almost anything a human mind can think of. Here is an example of the extension of the models to physical systems - The fluid-structure interaction of a bat's flexible wing. Engineers, for a while, tried to understand the aerodynamics of a bat flight and draw inspiration for advanced aerial systems but, due to challenges in the modeling methods, were unable to produce anything significant.

These are some of the earliest and most realistic simulations in this area. These simulations were performed on ARCH's Rockfish supercomputer using 2000+ CPU cores.