Hariprashad Ravikumar

Graduate Research Assistant - Machine Learning & Hpc @New Mexico State University

Las Cruces, NM, US

Engineered a complete end-to-end ML pipeline using Symbolic Regression (PySR) to analyze high-dimensional observables, achieving 93%+ model fit accuracy and successfully extracting novel, interpretable analytical functions from complex simulation data.+ Accelerated a multi-terabyte data processing workflow by 10x by developing a GPU-accelerated pipeline using C++ and CUDA (cuFFT) on high-performance computing (HPC) clusters.+ Ensured model reliability and numerical stability by creating production-grade Python packages for rigorous uncertainty quantification, implementing jackknife and bootstrap resampling across correlated configurations.+ Developed custom, physics-constrained loss functions in PyTorch/TensorFlow, significantly improving model generalizability and ensuring predictions adhered to domain-specific constraints.+ Designed and deployed custom SLURM orchestration workflows to manage and execute over CPU/GPU compute hours on the NERSC Perlmutter supercomputer, enabling robust, automated parallel analysis for large-scale projects.

New Mexico State University

Doctor of Philosophy - PhD, Physics

National Institute of Technology, Jalandhar, Punjab, India

Master of Science, Physics

New Mexico State University

Master of Science - MS, Physics