Latest
Quantum dynamics in symmetry-breaking states of correlated electrons: Antiferromagnetic phase
Physica Scripta 101, 155904 (2026)journal
Mukesh Khanore
Computational Physicist & PhD Researcher
I build machine-learning interatomic potentials that make large-scale atomistic simulations of ferroelectric perovskites run orders of magnitude faster while keeping ab-initio accuracy. My work spans model training, scientific software development and high-performance computing on national and European supercomputers.
I am a computational physicist finishing a PhD at Charles University and the Institute of Physics of the Czech Academy of Sciences. My research trains machine-learned interatomic potentials on ab-initio data so that simulations which would take weeks of DFT can run in hours without giving up quantum-level accuracy.
Before returning to academia I spent two years in industry as a data scientist and analyst — leading small teams, building forecasting models for electricity demand and renewable generation, and turning messy operational data into something decision-makers could act on.
Day to day I work in Python and FORTRAN on Slurm clusters, with a background in numerical methods, sparse linear algebra and GPU workflows. I have published on strongly correlated electrons, ultracold atoms in optical lattices, and the energy landscapes of perovskites.
6 entries · 2014–2026
Publications
-
2026
Deep learning frameworks for exploration of energy landscapes of perovskites
-
2025
Transition metals: Matching microscopic quantum dynamics with macroscopic magnetic order
-
2022
The quantum vortex states in extended Bose–Hubbard model: effects of lattice geometries, inter-particle interactions and spatial inhomogeneity
-
2015
Coexistence of Mott and superfluid domain of bosons confined in optical lattice
-
2014
Bose–Hubbard model with attractive interactions and inhomogeneous lattice potential
Experience
-
Oct 2022 – Present
PhD Researcher
Institute of Physics, Czech Academy of Sciences
- Developing a machine-learning approach to accelerate large-scale atomistic simulations of ferroelectric perovskites while retaining ab-initio accuracy.
- Fine-tune GRACE machine-learned interatomic potentials on curated VASP MLFF datasets; validate against DFT energies, forces, phonons and phase diagrams.
- Presented results at the 15th European Meeting on Ferroelectricity (Katowice, 2025), the XXVI Czech–Polish seminar (Milovy, 2026) and the E-MRS fall meeting (Warsaw, 2026).
-
Feb 2022 – Aug 2022
Senior Data Analyst
Sourcelink Software, Inc. (dba Dosii)
- Led the data analytics team; owned data extraction, cleaning and validation pipelines.
- Designed and applied predictive algorithms and presented insights to stakeholders and management.
-
Dec 2020 – Oct 2021
Deputy Manager (Data Science)
Regent Climate Connect Knowledge Solutions Pvt. Ltd.
- Managed the weather-forecasting team; responsible for error analysis and integration of data from multiple forecast providers.
- Contributed to machine-learning weather models for electricity demand and renewable (solar and wind) power forecasting.
-
Oct 2019 – May 2020
Consultant
Skymet Weather Services Pvt. Ltd.
- Developed a parameter-based mathematical model in FORTRAN for weather-optimal ship routing based on forecast data.
-
Aug 2013 – Aug 2019
Project Assistant
Department of Physics, S. P. Pune University
- Studied ground-state properties of ultracold atoms in rotating optical lattices via the extended Bose–Hubbard model, using both analytical and numerical methods.
- Implemented large sparse-matrix exact diagonalization with the Davidson algorithm for Hilbert spaces of dimension 10⁷.
Education
-
PhD in Physics
Charles University, Prague
Faculty of Mathematics and Physics.
-
M.Sc. in Physics
University of Pune
Department of Physics.
-
B.Sc. in Physics
S. P. College, University of Pune
Projects
PDF to BibTeX
Scans a folder of paper PDFs, extracts the DOI embedded in each one and resolves it into a formatted BibTeX entry. Built to keep a growing LaTeX bibliography in sync with a reference library without retyping citations by hand.
PythonLaTeXBibTeXDOI
Descriptors Visualization
Desktop GUI for generating structural descriptors from atomic configurations and plotting them interactively. Used to sanity-check the feature representations that go into machine-learned interatomic potentials before training.
PythonGUIDescriptorsMaterials Science
1D Hubbard Model
Numerical solver for the one-dimensional Hubbard model of correlated electrons, computing ground-state properties across the interaction strength range. A compact reference implementation of the exact-diagonalization approach used in my published work.
PythonExact DiagonalizationCondensed Matter
Qiskit Testing
Working notebooks exploring IBM's Qiskit framework — circuit construction, measurement and simulator backends — as a way into quantum algorithms for many-body problems.
QiskitPythonQuantum Circuits
Percolation Project
Monte Carlo study of percolation on lattices: locating the critical threshold, tracking cluster growth and visualising how connectivity emerges as site occupation increases.
PythonMonte CarloStatistical Mechanics
Map Plotting Python
Plotting utilities for geospatial forecast data, producing publication-quality maps of gridded fields. Written during my weather-forecasting work, where every model run needed a readable spatial view.
PythonMatplotlibGeospatial
Gridded Data
Routines for reading, regridding and subsetting grid-based scientific datasets, so that outputs from different forecast providers on different grids can be compared on common ground.
PythonNumPyData Pipelines
Any Base Number Representation
Small library for converting numbers between arbitrary bases in both directions, including fractional parts — the kind of primitive that turns up repeatedly in encoding and indexing problems.
PythonAlgorithmsNumber Theory
Remote Access
Helper scripts for working on remote compute clusters: opening connections, moving job files and retrieving results, cutting the boilerplate out of a daily HPC workflow.
ShellSSHHPC
Skills
Programming
- Python
- FORTRAN
- C
- Julia
- SQL
- Mathematica
- LaTeX
- gnuplot
- Qiskit
Simulation & Tools
- VASP (DFT, MLFF)
- LAMMPS
- ASE
- Phonopy
- GRACE / ACE-type MLIPs
- TensorFlow
- scikit-learn
- Graph neural networks
- Slurm / PBS clusters
- Singularity / Apptainer
- Git
- Linux
- Google Cloud
Research & Physics
- Density functional theory
- Machine-learned interatomic potentials
- Ferroelectric perovskites
- Strongly correlated electrons
- Ultracold atoms in optical lattices
- Exact diagonalization
- Phonon and phase-diagram calculations
- Numerical methods and sparse linear algebra
Conferences
XXVI Czech–Polish seminar
Presented “Deep Learning Frameworks for Exploration of Energy Landscapes of Perovskites”. Milovy, Czech Republic.
The 15th European Meeting on Ferroelectricity
Poster: “Machine-learning-based Investigation of the Energy Landscape of PbTiO₃, BaTiO₃ and CaTiO₃”. International Congress Centre (ICC), Katowice, Poland.
Workshop on Classical & Quantum Machine Learning for Condensed Matter Physics
Online workshop organized by ICTP Trieste, Italy.
e-INFRA CZ CONFERENCE
Hotel Occidental Praha, Prague.
Introductory School on Parallel Programming and Architecture
ICTP, Miramare, Trieste, Italy.
Teaching experience
Visiting Faculty
Taught Quantum Mechanics, Statistical Mechanics and Subatomic Physics to second-year B.Sc. students on the blended course.
Visiting Faculty
Delivered Basic Mechanical Engineering lectures and practicals to first-year B.Tech and M.Tech students, through December 2020.