Sourav Mondal

Hi, I'm Sourav

AI/ML Scientist | Computational Chemist | Drug Discovery

Computational scientist with 10+ years combining AI/ML with drug discovery and materials science. Currently at Aganitha Cognitive Solutions, building end-to-end pipelines for crystal structure prediction, conformer generation, pKa estimation, and ADMET property calculations. Published in Nature Computational Materials, JACS, and Nano Letters.

View Projects Download CV Publications Get in Touch

About Me

I'm an AI/ML Scientist specializing in drug discovery and computational chemistry with 10+ years of experience. Currently a Scientist at Aganitha Cognitive Solutions, I develop Graph Neural Networks for crystal structure prediction, AI models for conformer generation, pKa prediction, and ADMET property calculations for pharmaceutical applications.

With a PhD from JNCASR Bangalore and postdoctoral research at Trinity College Dublin, I've published in Nature Computational Materials, JACS, and Nano Letters. My expertise spans deep learning (PyTorch, TensorFlow, GNNs), quantum chemistry (DFT, molecular modeling), and cheminformatics (RDKit).

I build production-grade AI/ML systems for molecular property prediction, combining rigorous computational chemistry with modern deep learning to accelerate drug discovery. Proficient in Python, cloud technologies (Azure, Docker), and end-to-end ML pipelines from research to deployment.

Projects & Research

🔬 Current Work at Aganitha Cognitive Solutions

Building next-generation AI/ML tools for pharmaceutical research and drug discovery

Crystal Structure Prediction Pipeline

Graph Neural Networks for rapid polymorph screening of drug-like molecules

Conformer Generation

AI-driven platform enabling environment-aware conformer generation for advanced drug discovery applications

pKa Estimation

Two complementary modeling approaches—QM-based and GNN-based—for pKa estimation under non-aqueous conditions

ADMET Prediction

Physics-based QM platform for high-accuracy prediction of key thermodynamic properties: solubility, logP, and solvation energy

PyTorch Graph Neural Networks RDKit CP2K Docker Azure

🧬 Drug Discovery & Computational Chemistry

Drug Discovery Triage Platform

Full-Stack New!

Full-stack drug discovery triage application built with React, FastAPI, and RDKit. Features real-time ADMET property predictions, QED scoring, PAINS alerts, and interactive molecular visualization for rapid compound screening and lead optimization.

React + TypeScript FastAPI RDKit ADMET Prediction
  • Real-time ADMET property calculations (lipophilicity, solubility, permeability)
  • QED drug-likeness scoring and PAINS substructure alerts
  • Interactive 2D molecular structure visualization
  • Dockerized deployment with CI/CD pipeline
  • Production-ready with comprehensive error handling
View on GitHub → Live Demo →

Solubility Dual-Graph GNN

Graph Neural Networks New!

Dual-graph interaction GNN for predicting molecular solubility (logS) from solute-solvent SMILES pairs. Models explicit solute-solvent interactions via bidirectional cross-attention, trained on 100K+ pairs from BigSolDB 2.0. Achieves R² = 0.90.

PyTorch MPNN Cross-Attention RDKit MLflow
  • Dual-graph: separate MPNN encoders for solute & solvent
  • Bidirectional cross-attention for interaction modeling
  • 100,983 solute-solvent pairs (BigSolDB 2.0)
  • R² = 0.90, RMSE = 0.388, MAE = 0.290
View on GitHub →

🧠 Deep Learning & Neural Networks

PyTorch Deep Learning

Deep Learning

Comprehensive implementation of fully connected neural networks with PyTorch. Focus on model optimization, convergence analysis, and production-ready architectures.

PyTorch Neural Networks Model Optimization
  • PyTorch modeling and optimization techniques
  • Data balancing and sampling strategies
  • Model convergence analysis
  • Overfitting/underfitting prevention
View on GitHub →

Image Classification

Computer Vision

Image classification using pre-trained models and transfer learning. Demonstrates ability to leverage existing models for new tasks.

Transfer Learning Computer Vision CNN
  • Cats and Dogs image classification
  • Fine-tuning pre-trained networks
  • Computer vision pipeline development
View on GitHub →

📊 Machine Learning Fundamentals

ML Pipelines

Classical ML

Comprehensive ML pipelines covering both classification and regression. Demonstrates end-to-end ML workflow from data preprocessing to model evaluation.

Random Forest Scikit-learn Feature Engineering
  • Classification and regression problems
  • Feature scaling and dimensionality reduction
  • K-fold cross-validation
  • Data balancing and model evaluation
Classification → Regression →

Clustering Algorithms

Clustering

Implementation of clustering algorithms for pattern discovery and data exploration. Essential techniques for molecular similarity and compound clustering.

K-means Hierarchical Clustering
  • K-means clustering implementation
  • Hierarchical clustering methods
  • Cluster analysis and visualization
View on GitHub →

⚛️ Quantum Chemistry & Materials Science

Quantum Property Prediction

Materials ML

Neural network models for predicting quantum properties of materials, specifically Nitrogen Vacancy centers in diamond. Demonstrates ML application to quantum systems.

Neural Networks Quantum Properties Materials Science
  • Property prediction from structural features
  • Quantum mechanical property modeling
  • Feature engineering for materials science
View on GitHub →

Publications

Published in top-tier journals including Nature Computational Materials, JACS, and Nano Letters

2025

The spin phonon relaxation of single molecules magnet in the presence of strong exchange coupling

Sourav Mondal, Julia Netz et al.

ACS Cent. Sci. (2025)

Molecular Magnets Spin Relaxation Exchange Coupling
View Paper →
2023

Spin-phonon decoherence in solid-state paramagnetic defects from first principles

Sourav Mondal and A. Lunghi

npj Comput. Mat. (2023)

Quantum Computing First Principles Spin Dynamics
View Paper →
2022

Unravelling the contributions to spin-lattice relaxation in Kramers single-molecule magnets

Sourav Mondal and A. Lunghi

J. Am. Chem. Soc., 50, 22965 (2022)

Molecular Magnets Machine Learning Spin Relaxation
View Paper →
2022

Identification and Manipulation of Defects in Black Phosphorus

R. Harsh, Sourav Mondal*, D. Sharma, M. Bouatou, et al.

J. Phys. Chem. Lett, 13, 6276 (2022)

* Equally Contributed

2D Materials DFT STM
View Paper →
2020

Direct Observation of the Reduction of a Molecule on a Nitrogen pair in Doped Graphene

M. Bouatou, Sourav Mondal*, C. Chacon, F. Joucken, et al.

Nano Letters, 20, 6908 (2020)

* Equally Contributed

Graphene Molecular Adsorption STM Spectroscopy
View Paper →
2019

Importance of Epitaxial Strain at a Spin-Crossover Molecule Metal Interface

C. Fourmental, Sourav Mondal*, R. Banerjee, A. Bellec, et al.

J. Phys. Chem. Lett, 10, 4103 (2019)

* Equally Contributed

Spin Crossover Surface Science DFT
View Paper →

For a complete list of publications, visit my Google Scholar profile.

Experience & Education

Scientist

Jan 2024 - Present

Aganitha Cognitive Solutions - Hyderabad

  • Developing CSP pipelines with Graph Neural Networks for drug polymorph prediction
  • Building AI models for conformer generation, pKa prediction, and ADMET properties
  • Creating workflows for thermodynamics: solubility, logP, and solvation energy
  • Managing projects, client communications, and proposal writing

Computational Scientist & Deep Learning Engineer

Aug 2023 - Dec 2023

QpiVolta Technologies - Bangalore

  • Developed machine learning force fields for solid-state electrolytes
  • Modeled thermodynamic properties using molecular dynamics
  • Led contract research on complex biological systems simulation

Postdoctoral Researcher

Mar 2021 - Jul 2023

Trinity College Dublin - Ireland

  • Built neural network models for quantum property prediction
  • Developed computational methods for spin dynamics simulation
  • Published in Nature Computational Materials and JACS
  • Combined ab-initio methods with machine learning approaches

PhD in Computational Material Science

Jan 2015 - Feb 2021

JNCASR - Bangalore

  • Thesis: Tailoring Properties of 2D Systems via Molecular Adsorption and Defect Engineering
  • Employed DFT to study electronic and magnetic properties of 2D materials
  • Published in top-tier journals including Nano Letters and J. Phys. Chem. Letters
  • Received CEFIPRA grant for research visits to Paris

M.Sc in Chemistry

2012 - 2014

IIT Guwahati

  • Qualified NET CSIR-UGC Junior/Senior Research Fellowship

Skills & Expertise

🧬 AI/ML for Drug Discovery

Graph Neural Networks PyTorch TensorFlow Transformers AI-Agents Molecular ML ADMET Prediction

⚗️ Computational Chemistry

CP2K ORCA psi-4 PySCF DFT/SAPT ML Force Fields Docking

💊 Cheminformatics

RDKit ASE SMILES/SMARTS Conformer Generation pKa Prediction ADMET

🤖 Machine Learning

Neural Networks Random Forest SVM Clustering Cross-validation Hyperparameter Tuning

💻 Programming & Tools

Python (Expert) Fortran NumPy/Pandas Scikit-learn n8n Git

☁️ Cloud & DevOps

Docker Microsoft Azure CI/CD MLOps

📚 Research & Domain

Drug Discovery Materials Science Quantum Chemistry Scientific Computing Publications (Nature, JACS)

Let's Connect

Actively seeking AI/ML Scientist positions in Drug Discovery & Pharmaceutical Research

  • AI/ML Scientist - Drug Discovery - Graph Neural Networks, Molecular Property Prediction, CADD
  • Computational Chemist - AI/ML - Quantum Chemistry, Cheminformatics, DFT
  • Machine Learning Engineer - Pharma/Biotech - Deep Learning, Model Development, MLOps
  • Research Scientist - Materials/Drug Discovery - First-Principle Calculations, AI-driven Design

Why I'm a Strong Fit:

  • ✅ Current pharmaceutical AI experience at Aganitha Cognitive Solutions
  • ✅ Deep expertise in quantum chemistry, molecular modeling & cheminformatics
  • ✅ Proven ML skills: PyTorch, TensorFlow, GNNs, production pipelines
  • ✅ Published in Nature Computational Materials, JACS, Nano Letters
  • ✅ PhD + Postdoc with 10+ years in computational science

📍 Hyderabad, India
📧 souravchembwn@gmail.com | 📱 +919740851654

Feel free to reach out to discuss opportunities or collaborations!

GitHub LinkedIn Email