Sibasankar Panigrahy

About

Computational materials scientist working where statistical mechanics, molecular simulation and machine learning meet.

Portrait of Sibasankar Panigrahy
GitHub OpenReview

I’m a computational materials scientist at QpiAI. I did my PhD in Materials Science & Engineering at IIT Delhi with Prof. Divya Nayar in the Computational Soft Matter Lab. There I studied how functionalised gold nanoparticles behave in crowded, chemically complex environments, and built machine-learning models, trained on simulation data, that predict and design polymer–surface adhesion.

My approach is to get the physics right first: validate the force field against experiment, compute free energies carefully, and connect trajectories to solvation thermodynamics. Then I use that validated data to train models that generalise across chemical spaces far larger than simulation alone could cover.

Before the PhD I completed an M.Tech in Materials Engineering at the Defence Institute of Advanced Technology, Pune.

I write regularly in the lab notebook about methods, papers I’m reading, and what I’m working on.

Path

  1. Computational Materials Scientist

    QpiAI

  2. PhD, Materials Science & Engineering

    Indian Institute of Technology Delhi
    Thesis: Rational Design of Functionalized Nanomaterials for Biological Applications Using Molecular Dynamics and Machine Learning. Advisor: Prof. Divya Nayar, Computational Soft Matter Lab.

  3. M.Tech, Materials Engineering

    Defence Institute of Advanced Technology, Pune

Talks, posters & workshop papers

  1. Data-Driven Prediction of Polymer Adhesion on Heterogeneous Surfaces via Attention based Learning

    Poster · CECAM workshop

  2. Data-driven prediction of polymer surface adhesion using high-throughput MD and hybrid network models

    Workshop paper · AI4Mat Workshop, NeurIPS 2025

  3. Data-Driven Prediction of Polymer Adhesion on Heterogeneous Surfaces via Attention based Learning

    Poster · AIMS 2025

Methods & tools

Simulation

All-atom molecular dynamicsCoarse-grained (Kremer–Grest) modelsUmbrella sampling & PMFsPolarizable Drude models (GolP-CHARMM)Force-field reparameterisation (NBFIX)

Theory & analysis

Kirkwood–Buff solvation theoryPreferential interaction coefficientsSolvation entropy decompositionMonte Carlo Ising pattern generation

Machine learning

CNN / GRU / Transformer hybridsTandem networks for inverse designBoolean particle swarm optimisationHyperparameter search (Optuna)

Tools

GROMACSPython · NumPy · PyTorch · TensorFlowVMD · OVITOHPC (PADUM, IIT Kanpur)

PhD research supported by the National Supercomputing Mission (NSM), India, with computing on IIT Delhi's PADUM and IIT Kanpur's HPC facilities.