Codes and molecules are the things that essentially entertain me.
Research Fellow · Computational Chemistry · Enhanced Sampling · Scientific AI
I am a Senior Research Fellow and computational scientist working at the intersection of molecular simulation, statistical mechanics, and machine learning. My work focuses on developing algorithmic frameworks to extract kinetic observables from rare-event dynamics, building scientific software for enhanced sampling, and designing AI-native systems for molecular discovery.
Research is formalized curiosity. It is poking and prying with a purpose. — Zora Neale Hurston
|
Developing path-resolved frameworks for weighted ensemble (WE) simulations to efficiently sample rare-event pathways and extract unbiased kinetic observables from complex free energy landscapes. Designing collective variables that capture reaction coordinates and transition states in molecular kinetics, with applications to nucleation, protein folding, and conformational change. |
Building physics-informed neural networks and variational autoencoders for automated phase classification, order parameter discovery, and transferable solvation free energy prediction. Creating interactive visualization systems and AI-native presentation frameworks for computational science communication and reproducible research narratives. |
Direction-guided adaptive sampling for rare-event pathway generation
An algorithmic framework using ultrashort monitored trajectories to rapidly generate transition pathways in high-dimensional molecular systems. Integrates with OpenMM and PLUMED for seamless HPC deployment.
Python OpenMM PLUMED Enhanced Sampling JCTC 2025
Variational autoencoder for automated ice polymorph identification
A deep learning system that identifies and classifies ice polymorphs directly from molecular simulation trajectories without manual feature engineering.
PyTorch VAE Molecular Dynamics Phase Classification
Open-source framework for weighted ensemble simulation analysis
A modular Python ecosystem for post-processing weighted ensemble trajectories, computing rate constants, constructing Markov state models, and visualizing path ensembles.
Python NumPy SciPy MDAnalysis HPC
Transferable solvation free energy prediction via graph neural networks
Comparative study of molecular representations and geometric deep learning methods for building accurate yet efficient implicit solvent models.
PyTorch Geometric GNNs Free Energy MBAR
| Year | Title | Venue |
|---|---|---|
| 2025 | PathGennie: Direction-Guided Adaptive Sampling for Rare-Event Pathways | J. Chem. Theory Comput. |
| 2025 | IceCoder: Automated Ice Polymorph Classification with Variational Autoencoders | J. Chem. Theory Comput. |
| 2023 | Computational Study of Efficient Light Harvesting in Self-Assembled Organic Luminescent Nanotubes | Chem. Sci. |
| In Prep. | Path-Resolved Free Energy Frameworks for Weighted Ensemble Simulations | Manuscript in Preparation |
| In Prep. | Graph Neural Networks for Transferable Solvation Free Energy Prediction | Manuscript in Preparation |
- Enhanced Sampling Methods
- Weighted Ensemble Simulations
- Molecular Kinetics
- Free Energy Calculations
- Scientific Machine Learning
- Computational Biophysics
- Scientific Visualization
- Research Software Engineering
- AI-Native Scientific Systems
- HPC Workflow Design