Computational Scientist at Idaho National Laboratory

Building computational tools for advanced nuclear systems.

I develop thermodynamically informed multiphysics methods for nuclear fuels, materials, and reactor systems. My work sits at the intersection of scientific computing, physics-based modeling, and research software development.

  • Multiphysics modeling
  • Advanced nuclear
  • Computational thermochemistry
  • Machine learning

Research

A systems view of nuclear modeling.

The site is organized around the work itself: computational thermochemistry, multiphysics coupling, advanced reactor analysis, and the software needed to make those models usable at scale.

Theme 01

Thermodynamics-coupled multiphysics

Developing methods that connect chemistry, transport, corrosion, and material behavior instead of treating them as isolated pieces.

Theme 02

Molten-salt reactor chemistry and corrosion

Modeling chemically evolving reactor environments across microstructural and engineering scales, with particular attention to species tracking and safety.

Theme 03

Reduced-order and machine-learning acceleration

Building surrogate strategies that preserve physical meaning while cutting the cost of high-fidelity thermodynamics-informed simulations.

Selected work

From core methods to deployable scientific tools.

Yellowjacket

During his doctoral work, Parikshit led development of a MOOSE-based Gibbs-energy minimizer for thermochemical equilibrium calculations.

  • Thermochemical equilibrium
  • MOOSE framework
  • Scientific software design

MSR species accountancy and depletion

Contributing to full-loop molten-salt reactor chemistry and corrosion modeling with coupled species accounting across multiphysics workflows.

  • Molten-salt systems
  • Corrosion modeling
  • Coupled analyses

Graphite property modeling

Working on material-property and mechanistic models for advanced nuclear-grade graphite, including uncertainty-aware calibration at high temperature.

  • Bayesian calibration
  • Materials degradation
  • Mechanistic modeling

Surrogate models for multiphysics simulation

Leading work on computational methods that accelerate thermodynamically informed reactor and materials simulations without discarding the governing physics.

  • Reduced-order models
  • Machine learning
  • Scalable simulation

Software

Code is part of the research argument.

The software layer is not secondary here. It is how the methods become testable, reusable, and useful to other researchers and engineering teams.

Framework

MOOSE

Multiphysics environment used across several strands of this work, from thermodynamics coupling to reactor chemistry applications.

Open repository

Thermochemistry tooling

Thermochimica and related equilibrium methods

Computational thermodynamics infrastructure for chemical state estimation, phase equilibrium, and constitutive-property workflows.

Open repository

Public profile

GitHub

Public code, contributions, and technical trail across scientific computing and open research software.

Visit profile

Publications

Recent work in public view.

For the complete and up-to-date publication record, visit Google Scholar.

Trajectory

An international path through mechanics, nuclear engineering, and computational science.

Canada

Doctorate in modelling and computational science

Doctoral research at Ontario Tech University focused on algorithms for thermochemical-equilibrium and MOOSE-based solver Yellowjacket.

Italy

Master's in nuclear engineering

Research at Politecnico di Milano included computational modeling of helium bubble reactivity feedback in molten-salt fast reactor systems.

India

Bachelor's in mechanical engineering

Undergraduate training at Dr. A.P.J. Abdul Kalam Technical Univeristy with capstone project on CFD modeling of exhaust manifold to improve internal combustion engine efficiency.

United States

Computational scientist at Idaho National Laboratory

Current work spans reactor chemistry, advanced materials, surrogate methods, and multiphysics software for nuclear applications.

Connect

Open to technical conversations that go beyond surface-level AI and modeling talk.

The clearest routes in are public: research profile, code, and professional network.