I bridge physics-based simulations and cheminformatics with machine learning to accelerate molecular design and drug discovery. My background spans fluid mechanics, polymer physics, and generative AI — from deriving continuum equations to training diffusion models on molecular data.
I’m open to collaboration across AI for Science, molecular modeling, and generative chemistry. Let’s connect and build the next generation of molecular design together.
🔗 Website · LinkedIn · Google Scholar · arXiv · ORCID
💊 Senior AI/ML Scientist @ Merck & Co. Applying predictive deep learning models and generative AI to accelerate drug discovery by designing, predicting, and optimizing small-molecule therapeutics.
🎓 Ph.D. Chemical Engineering, Caltech (2024) — Wang & Brady Groups 🎓 M.S. Chemical Engineering, Caltech (2022) 🎓 B.S. Chemical Engineering, UC Berkeley (2019)
| Repository | Description |
|---|---|
| OpenADMET-ExpansionRx | ADMET stacked ensemble models — 18th / 400+ in Polaris blind challenge |
| DDPM-Enhanced-Sampling | Denoising diffusion models for Boltzmann-consistent polymer conformational sampling |
| Swimming-in-Potential-Flow | C++/CUDA boundary integral methods — companion code for JFM 2022 paper |
- Binding Modes and Water-Mediation of Polyelectrolyte Adsorption to a Neutral Calcium Carbonate Surface — Langmuir, 2025 DOI
- Multivalent Ion-Mediated Polyelectrolyte Association and Structure — Macromolecules, 2024
DOI, arXiv - Adsorption Isotherm and Mechanism of Calcium-Ion Binding to Polyelectrolyte — Langmuir, 2024
DOI, arXiv - Swimming in Potential Flow — J. Fluid Mech., 2022
DOI - Geometry and Dynamics of Lipid Membranes: The Scriven–Love Number — Phys. Rev. E, 2020
DOI, arXiv
Core Areas: Graph neural networks (GCNs, MPNNs, GATs) · Generative models (VAEs, diffusion, GFlowNets, transformers + RL) · Uncertainty quantification · Transfer & curriculum learning · Conformer ensemble models
- Drug Discovery: Structure-based & ligand-based design, de novo generation, lead optimization, multi-parameter optimization (MPO), ADMET prediction, synthesizability screening, virtual screening
- Representations: SMILES, InChI, molecular graphs, 3D conformers, ECFP fingerprints, pharmacophores, pre-trained embeddings (ChemBERTa, CheMeleon)
- Databases & Tools: ChEMBL, PubChem, DrugBank; Schrödinger Suite, AutoDock Vina
- Molecular Dynamics: Enhanced sampling (metadynamics, OPES, umbrella sampling, HREX), MDAnalysis, Markov state modeling, free-energy calculations
- Continuum & Quantum: Potential flow theory, microswimmer hydrodynamics, lipid membrane mechanics (Scriven–Love); ORCA for QM; Schrödinger/GLIDE for docking
Python · C++ · CUDA · SQL · Shell · LaTeX · FORTRAN
Slurm · PBS · Ansible · Spack · Git
📄 For more on my research, experience, and publications, visit alec-glisman.github.io