Computational Materials • AI/ML • Quantum Algorithms

Quantum simulation meets materials science

I'm a computational materials scientist building the bridge between first-principles physics, machine learning, and near-term quantum algorithms — with the goal of making molecular and materials simulation both faster and more honest about its own limits.

My work spans DFT-driven ML for battery materials to VQE/SQD-based quantum simulation of real large molecules. What ties it together is a focus on reproducibility, interpretability, and workflows that actually run on the hardware that exists today.

View research projects View CV

First-principles simulation

• DFT & AIMD (VASP, QE)
• Structure-property relationships
• High-throughput workflows

Materials-focused ML

• Graph Neural Networks (M3GNet)
• Physics-informed descriptors
• Active learning for discovery

Quantum algorithms

• VQE / qEOM for excited states
• Spectral filtering methods
• NISQ-era chemical simulations

Current Research Directions

Spectral filtering for correlated systems. Developing quantum algorithms that isolate near-degenerate eigenvalues in molecular Hamiltonians — targeting noise robustness and scalability on NISQ hardware beyond what VQE alone can reach.
Quantum PCA via spectral gap amplification. Applying quantum principal component analysis to molecular density matrices, with a goal of exponential circuit-depth compression. Manuscripts in preparation; code under active development.
Active-space design for bio molecules. Extending the VQE/qEOM/SQD framework from LiPF₆ and NaPF₆ to larger bio-molecular systems, benchmarking against CCSD(T) to establish where quantum methods offer genuine advantage.

I'm interested in the problems that sit exactly at the boundary of what classical simulation can handle — where DFT becomes qualitatively wrong and quantum methods become the only honest option.

Conference & Seminar Activity

Korea-India Research Joint Workshop, Seoul 2025

International Conference on Sustainable Batteries, Kolkata 2025

International School on Modelling and Simulations, JNCASR 2025

Winter School on Electronic Structure, Kathmandu University 2020

Machine Learning in Materials Science seminar, IISER Pune 2018

Selected conference and seminar photographs.

About

Computational materials scientist working at the intersection of first-principles simulation, machine learning, and near-term quantum algorithms — with a focus on problems that are chemically relevant, not just computationally convenient.

Background & motivation

I grew up in West Bengal and studied physics through a period when "quantum computing" still felt like science fiction. By the time I finished my PhD at IISER Pune — working on antiperovskite battery materials — the gap between what quantum hardware could do and what materials science needed had narrowed enough to take seriously.

That convergence is what drives my work. I'm interested in the problems that sit exactly at the boundary of what classical simulation can handle: correlated electronic states, excited-state spectra, and molecular systems where DFT becomes qualitatively wrong and quantum methods become the only honest option.

PhD, IISER Pune Postdoc, IKST Bengaluru JRF, IIT Bombay

Research focus

My work spans two interconnected tracks. On the materials side, I build DFT- and AIMD-driven machine learning pipelines for predicting electrochemical properties of battery materials — combining physics-informed descriptors with graph neural networks to make prediction both accurate and interpretable.

On the quantum side, I develop and benchmark variational and subspace-based quantum algorithms — VQE, qEOM, SQD, and spectral filtering methods — applied to real molecular systems such as battery electrolyte salts (LiPF₆, NaPF₆, FSI). The emphasis is on active-space design, excited-state accuracy, and dissociation behaviour on near-term hardware.

First-principles simulation Materials-focused ML VQE · qEOM · SQD

Methodological approach

I design end-to-end research pipelines spanning simulation, data extraction, model construction, and validation. What connects all of them is a commitment to reproducibility — not as a formality, but because most computational results in our field are quietly unreproducible. Every workflow I publish comes with the code, the parameters, and an honest account of where the method breaks down.

In my quantum computing work specifically, I focus on spectral isolation and subspace compression for chemistry problems, rather than hardware benchmarking for its own sake. The question I keep returning to is: what can a quantum algorithm do for a real chemical system that a classical method genuinely cannot?

Reproducible workflows Spectral methods Model interpretability

Selected achievements

Three first-author publications in 2026 in Advanced Quantum Technologies, Small Methods and JPCC with 9 peer-reviewed journal articles spanning quantum simulation, battery materials,and materials AI/ML.

Recipient of the MHRD doctoral fellowship (2016) and teaching assistant award for the Winter School on Electronic Structure and Molecular Dynamics Simulations (2020). Certified in Quantum Computing jointly by C-DAC and IIT Roorkee (2025).

9 journal articles MHRD fellowship QC certified 2025