PhD student in Bioengineering, University of Washington · Imoukhuede Lab · expected 2027

I build computational models of oxytocin receptor (OXTR) signaling, from literature-scale parameter synthesis through mechanistic simulation to generative ligand design.

• Calibrated uncertainty for small-data property prediction. GNN ensembles with temperature scaling, isotonic CDF recalibration, scaffold-based splits, CRPS and interval coverage. Code: oxtr-gnn-calibration
• Kinetic meta-analysis of Class A GPCR signaling. Bayesian hierarchical (PyMC) and REML random-effects (metafor) pooling of rate constants across 57 studies, with heterogeneity quantification and simulation-based power analysis. Curated dataset and analysis code release with the manuscript.
• Mechanistic ODE modeling of receptor signaling. Mass-action models of OXTR-Gq membrane dynamics in Tellurium and SimBiology, global sensitivity analysis, identifiability triage, and model selection by AIC/BIC and likelihood ratio tests.
• Biased agonist design. DrugEx graph transformer fine-tuning, multi-objective reinforcement learning, QSPR reward models, and dual-state docking, run as a checkpointed Snakemake workflow under SLURM.
• Reproducible scientific computing. Making all of the above run end to end on someone else's machine. Worked example: recapitulating a published GPCR model

Extending the generative work toward pathway-selective ligand design, and moving three computational OXTR projects toward publication.

Systems-level modeling of biological systems, especially GPCR pharmacology and signal transduction. Also happy to talk with anyone building parameter databases or doing meta-analysis in pharmacology. That tooling is badly underbuilt.

• Modeling & inference Python NumPy/SciPy PyMC pandas R (metafor) Tellurium MATLAB SimBiology
• ML PyTorch scikit-learn GNNs DrugEx QSPR SALib
• Cheminformatics & structure RDKit GNINA
• Infrastructure Docker/Apptainer SLURM (UW Hyak) Snakemake Git WSL Conda

LinkedIn · Email · Google Scholar