I’m a bioinformatician and AI research scientist at Stanford University School of Medicine, working where deep learning, molecular modeling, and scientific software meet. My research focuses on AI-driven protein design and computational drug discovery—including safer anesthetic candidates for field and battlefield medicine.
I build end-to-end research systems: from foundation models, graph neural networks, and molecular simulation to reproducible pipelines, evaluation frameworks, and tools scientists can actually use. Over 8+ years, my work has crossed computational biology, clinical genomics, high-performance computing, and open-source engineering.
Research north star: compress the distance between a promising biological idea and a testable therapeutic hypothesis.
- Generative and predictive models for protein structure and function
- Structure-based drug design, molecular docking, and ADMET intelligence
- Biomedical knowledge graphs, RAG, and evidence-grounded scientific agents
- Reproducible, scalable pipelines for multi-omics and high-throughput discovery
- Honest benchmarks that distinguish model performance from scientific utility
|
Physics-based molecular docking with flexible-ligand search and SE(3)-equivariant GNN rescoring. Research signal: Pearson R = 0.88 on the PDBbind refined set; reproducible search, rich interaction analysis, and publication-ready reports. 
|
Protein AND ligAnd interaction MAPper — comprehensive detection, visualization, and empirical binding affinity estimation for protein–ligand complexes. Design goal: protein-ligand interactions visualization tool. 
|
|
Graph neural networks predict molecular ADMET properties while a RAG-based explainer connects predictions to mechanistic context.
|
R-GCNs and biomedical knowledge graphs uncover drug–disease relationships, paired with evidence-aware natural-language explanations.
|
|
Deep-learning clinical decision support for predicting cancer cell-line drug response with context-specific RAG explanations. |
Tracks and analyzes PubMed and bioRxiv literature for drug targets, compounds, and therapeutic areas. |
flowchart LR
A[Biological question] --> B[Curated data]
B --> C[Representation learning]
C --> D[Predictive or generative model]
D --> E[Physics-aware validation]
E --> F[Reproducible workflow]
F --> G[Testable therapeutic hypothesis]
G -. experimental evidence .-> A
AI & scientific computing
Scientific workflows & infrastructure
Molecular & biomedical systems
Deep-dive metrics dashboard
Contribution map
- Published work spanning AI-enabled drug design, molecular modeling, clinical genomics, nanotherapeutics, and biomolecular simulation.
- Research experience across Stanford University, Uppsala University, Karolinska Institutet, DKFZ, and University Medical Center Freiburg.
- Nextflow Ambassador and Sigma Xi member; builder of educational resources in AI-driven drug discovery, NGS, quantum chemistry, and molecular dynamics.
- Recipient of research and innovation support from Colgate-Palmolive, Uppsala University/ABB–Hitachi, and Karolinska Institutet.
I’m open to research collaborations, translational partnerships, and open-source work in:
- AI-driven protein and therapeutic design
- Computational biology and multi-omics infrastructure
- Foundation models and knowledge systems for biology
- Reproducible scientific software and benchmarking
Science moves faster when the tools, evidence, and ideas are open.