Postdoctoral Researcher · IISc Bengaluru

Manas
Sharma

PhD in Physics · Computational Materials Scientist

I develop electronic structure methods such as quantum embedding, machine-learned interatomic potentials, and computational codes to understand light-matter interactions, thin-film growth, and the electronic structure of strongly correlated and complex materials — at the atomic scale.

View Publications Download CV ↗
Manas Sharma
14 Publications
491+ Citations
8 h-index

Updates

News

Jul 17, 2026

NPDF awarded

Awarded an NPDF for a project on the ML-accelerated study of MoS2 thin-film deposition.

Jul 8, 2026

MLIP Studio launched

MLIP Studio, an open platform for interactive MLIP benchmarking and atomistic simulations, launched alongside its arXiv preprint.

Jan 01, 2026

PyFock released

PyFock, a pure-Python MIT-licensed DFT code with molecular integrals and exchange-correlation terms, was released.

Jul 2025

Phys Whiz crossed 20k subscribers

Phys Whiz reached 20k subscribers with tutorials and lectures on physics, computational materials science, and numerical methods.

May 2024

Cube Suite released

Cube Suite, a web app for processing and manipulating CUBE files, was released.

Mar 2024

PhD thesis defended

Defended the PhD thesis with the summa cum laude grade.

Jul 2023

RIPER-TOOLS released

RIPER-TOOLS launched for creating TURBOMOLE RIPER input files from CIF, XYZ, POSCAR, Materials Project, and PubChem sources.

Background

About Me

I am a postdoctoral researcher in the Chemical Engineering Department at the Indian Institute of Science (IISc), jointly advised by Prof. Ananth Govind Rajan and Prof. Sudeep Punnathanam. My current work focuses on developing graph neural network (GNN)-based interatomic potentials for simulating thin-film growth processes such as CVD and PVD.

Previously, I obtained my Ph.D. (Physics) from Friedrich Schiller University Jena, Germany, with the highest distinction of summa cum laude, under Prof. Dr. Marek Sierka. During my doctorate I developed and implemented advanced DFT-based quantum embedding methods coupled with wavefunction (MP2, CCSD), many-body (GW/BSE) and real-time TDDFT approaches, contributing new capabilities to the TURBOMOLE quantum chemistry package.

Beyond research, I am deeply invested in scientific outreach: I run the Phys Whiz YouTube channel (~21k subscribers, 3.6M views), build open-source computational tools, and mentor students worldwide.

DFT RT-TDDFT Quantum Embedding GNN Potentials MLIPs CVD / PVD NEB TURBOMOLE PySCF VASP Python Fortran

Quick Info

Focus Areas

Research Interests

Multiscale Methods and Quantum Computing

Quantum Embedding Methods

Developing DFT-based embedding methods coupled with quantum computing algorithms like variational quantum eigensolver (VQE), correlated wavefunction methods (MP2, CCSD), many-body Green's function methods (GW/BSE) and RT-TDDFT to describe strongly correlated systems and study high-accuracy excited-state and optical properties at reduced cost.

MLIPs

Machine-Learned Potentials

Training, finetuning, distilling graph neural network (GNN)-based machine learning interatomic potentials (MLIPs) on in-house generated high-fidelity DFT data to enable large-scale atomistic simulations of thin-film deposition (CVD/PVD). Experience gained in the development of electronic structure methods and codes is leveraged in generating high-accuracy data efficiently utilizing. Also working on the generation of foundational datasets for universal MLIPs

RT-TDDFT

Light-Matter Interactions

Studying high harmonic generation and optical excitations in complex materials using real-time TDDFT and GW/BSE approaches facilitated by self-developed methods and codes.

Dynamics

Thin-Film Growth

Investigating reaction pathways and energy barriers for CVD/PVD using DFT, nudged elastic band (NEB), and ab initio molecular dynamics (AIMD) simulations. Additionally, the data generated from DFT is utilized to train/finetune MLIPs to run molecular dynamics and see thin-film growth in real-time, offering unprecedented insight into nucleation and growth.

Software and Tools

Computational Code Development

Building performant DFT codes, neural network libraries, and GUI applications — optimized for parallelization and GPU acceleration — to serve the research community.

Dissemination

Scientific Outreach

Creating YouTube tutorials, web and Android apps, and blog content to make computational materials science accessible to students and researchers worldwide.

Trajectory

Education & Experience

2024 - Present

Postdoctoral Researcher · IISc Bangalore

Developing GNN-based MLIPs for thin-film deposition and materials design, with DFT, NEB, AIMD, and molecular dynamics workflows.

2019 - 2024

PhD Physics · Friedrich Schiller University Jena

Summa cum laude. Thesis on density functional theory based embedding for molecular and periodic systems.

2019 - 2024

Scientific Employee · FSU Jena

Implemented quantum embedding and real-time TDDFT capabilities in the TURBOMOLE RIPER module using Fortran and Python.

2016 - 2018

MSc Physics · University of Delhi

Specialized in nanoscience, ranked third, and completed DFT-based research that led to publications.

2013 - 2016

BSc (Hons.) Physics · University of Delhi

Acharya Narendra Dev College. Ranked second with 87.08 percent and received the DC Arora scholarship.

2014 - Present

BragitOff & Phys Whiz

Writing tutorials, building apps, and creating physics outreach content through BragitOff.com and Phys Whiz.

Recent Work

Selected Publications

All Publications →   Google Scholar ↗

Code Development

Tools & Applications

Electronic Structure

TURBOMOLE

Electronic-structure package to which I contributed quantum embedding, real-time TDDFT, and related methods for periodic systems.

Project site

Python DFT

PyFock

Pure-Python DFT code with molecular integrals and exchange-correlation terms.

Project site

MLIP Web App

MLIP Studio

Open, no-code platform for running, testing, and comparing more than 60 machine-learned interatomic potentials.

Launch app

Quantum Chemistry

RIPER-Tools

Web tools for generating TURBOMOLE RIPER inputs and processing or manipulating volumetric CUBE files.

Open tool

Scientific Applications

CrysX

Android, desktop, and web applications for crystallography, molecular and crystal visualization, augmented reality, and computational chemistry.

Explore apps

VASP Web App

VASP-GUI

Online GUI for preparing VASP inputs and parsing or visualizing VASP output files.

Project page

Get in Touch

Let's Connect

Whether it's research collaboration, outreach, or just a conversation about computational physics — I'd love to hear from you.

[email protected]
GitHub LinkedIn Twitter / X YouTube