Research Portfolio

Atomistic Intelligence for Sustainable Chemical Engineering

Developing next-generation catalysts through computational design and machine learning

Research Mission

My research leverages machine learning, multiscale simulations, and ab-initio calculations to discover efficient catalytic materials for clean fuel production and sustainable chemical processes. I focus on bridging the gap between theoretical predictions and experimental observations through data-driven methods.

Core Research Areas

Catalyst Discovery

"It doesn't matter how beautiful your theory is, it doesn't matter how smart you are. If it doesn't agree with experiment, it's wrong." - Richard Feynman

Active learning accelerates materials screening through intelligent sampling.

CEAL Framework

Catalyst Embedding Active Learning - A groundbreaking framework that replaces brute-force screening with targeted sampling guided by pretrained GNN models. Uses latent space similarity and predictive uncertainty to enable efficient identification of promising candidates across complex chemical spaces.

Key Methodologies
Graph Neural Networks Active Learning Uncertainty Quantification Latent Space Analysis Chemical Space Mapping

Multiscale Simulation

Bridging atomic and macroscopic scales

Graph neural networks enable atomic to nanoscale modeling. [Chem. Eng. J. 2024]

CRACK Framework

Contrastive Relational-Aware Compression of Knowledge - A novel distillation framework using contrastive learning that compresses large GNN potentials into lightweight surrogates, achieving over 50× speedup in energy and force prediction while maintaining accuracy.

Key Methodologies
Knowledge Distillation Contrastive Learning Graph Compression Transfer Learning Model Optimization

Experimental Realization

Validating computational predictions

Theory-guided synthesis validates computational predictions. [Angew. Chem. 2025, Adv. Energy Mater. 2022]

System Types Studied
  • Metal-hydroxide heterostructures
  • Nanostructured ceria systems
  • Single atom Pt on ceria
  • Ni exsolution structures
Applications
  • CO₂ hydrogenation
  • Oxygen reduction reactions
  • Methane oxidation
  • Hydrogen evolution
Key Methodologies
Multiscale Modeling Molecular Dynamics DFT Calculations Kinetic Monte Carlo Machine Learning Potentials