Computational Chemist & Postdoctoral Fellow · University of Toronto / Acceleration Consortium
I am a computational chemist who develops and applies quantum mechanical methods for molecular science. My work spans density functional theory, dispersion interactions, and machine learning for molecular property prediction. Named developer for FHI-aims. Currently working with Prof. O. Anatole von Lilienfeld at the University of Toronto.
720+ citations · h-index 14 · 18 publications · 10 invited talks · $10M+ grant funding
My research is rooted in the fundamental physics of molecular interactions and building better computational tools for chemistry.
Dispersion & Density Functional Theory — I develop and implement methods for accurately treating London dispersion interactions in DFT. This includes the exchange-hole dipole moment (XDM) model in Psi4, adaptive hybrid functionals (aPBE0), non-local adiabatic connection methods (nLanE), and double-hybrid functionals (DH24) in FHI-aims. I am currently generating training data for a machine-learned version of XDM dispersion following Tu et al..
Machine Learning Interatomic Potentials — I develop physics-constrained approaches to learning interatomic potentials, where the model learns universal scaling parameters for known functional forms rather than fitting the potential surface directly. I also fine-tune MACE neural network potentials for targeted chemical domains, including a pipeline for 1,445 pesticide molecules trained on wB97M-D3BJ/def2-TZVPPD data.
D. Khan, A. J. A. Price, M. L. Ach, O. A. von Lilienfeld. Adaptive hybrid density functionals. Science Advances 2025, 11(5), eadt7769.
A. J. A. Price, A. Otero de la Roza, E. R. Johnson. XDM-corrected hybrid DFT with numerical atomic orbitals predicts molecular crystal energetics with unprecedented accuracy. Chem. Sci. 2023, 14, 1252.
A. J. A. Price, K. R. Bryenton, E. R. Johnson. Requirements for an accurate dispersion-corrected density functional. J. Chem. Phys. 2021, 154, 230902. (Top 10% Altmetric)
M. DeJong, A. J. A. Price et al. Small molecule binding to surface-supported single-site transition-metal reaction centres. Nat. Commun. 2022, 13, 1-10.
| Repository | Description |
|---|---|
psi4_xdm |
XDM dispersion + aPBE0 + nLanE adaptive DFT in Psi4 (C++/Python) |
adaptive-ml-potentials |
Physics-constrained parameter learning for interatomic potentials (Price & von Lilienfeld, 2026) |
molecular-ml-pipeline |
MACE fine-tuning pipeline: 28K pesticide structures, MLflow tracking, GPU training |
agentic-ai-demos |
Agentic AI for drug discovery: LangGraph agents, molecular RAG, PubChem integration |
teaching |
DFT lectures for CHM328 at UofT (Beamer slides + companion notes) |
cv |
Academic CV with GitHub Actions auto-build |
| Course | Role | Institution |
|---|---|---|
| CHM328: Modern Physical Chemistry | Guest Lecturer (DFT & Computational Chemistry) | University of Toronto, Winter 2026 |
| Domain | |
|---|---|
| Quantum Chemistry | Psi4 (developer), FHI-aims (named developer), Gaussian, critic2, postg, ASE |
| Theory | DFT, XDM dispersion, CCSD(T), adiabatic connection, double hybrids, perturbation theory |
| Machine Learning | PyTorch, MACE, scikit-learn, KRR, cMBDF molecular descriptors |
| Scientific Computing | Fortran, C++, Python, MPI/OpenMP, SLURM, Alliance Canada HPC, Globus |
| Infrastructure | Proxmox, Docker, Ansible, Tailscale, Prometheus/Grafana, ZFS, MLflow |
- 2026 — ML-XDM dispersion model: large-scale DFT data generation for machine-learned dispersion corrections
- 2026 — Collaborator on $10M NRF Singapore-UofT grant for ML design of complex materials
- 2025 — Invited seminars at Seoul National University, Fritz-Haber-Institut (Max Planck), University of Cambridge, National University of Singapore
- 2025 — Science Advances publication on adaptive hybrid density functionals
- 2022 — Best graduate student poster, Canadian Symposium on Theoretical & Computational Chemistry