Projects

Scientific software and computational systems.

Selected projects emphasize reproducibility, simulation, data pipelines, and engineering use—not generic application development.

Scientific simulation

Agentic APSIM

Research software around APSIM for simulation configuration, experiment planning, execution, qualification, provenance, and evidence-supported analysis. The design focuses on retaining model validity and traceability while reducing manual workflow overhead.

PythonAPSIMValidationResearch software
Building physics

Classroom airflow & geometry pipeline

End-to-end research workflow connecting captured indoor geometry, reconstruction and cleanup, CFD meshing, airflow solution, and contaminant transport for real classroom environments.

OpenFOAMCFDHPC3D reconstruction
Scientific machine learning

PhySiViT

A physics-simulation vision transformer trained on simulation data and evaluated against general-purpose vision foundation models for classification, temporal forecasting, and physics clustering.

Vision transformersThe WellCerebras
3D reconstruction

Plant NeRFs

Mobile-image and pose data, LiDAR ground truth, and evaluation workflows for plant geometry reconstruction with Neural Radiance Fields. The work supports the 2024 Plant Phenomics publication.

NeRFLiDARComputer vision
Grid analysis

Grid-Gent

Lightweight distribution-grid scenario screening for load, solar, and hosting-capacity what-if analysis with support for uploaded feeder data and transparent reasoning.

PythonFastAPIPower systems
Environmental data

Heat, Smoke & Grid Reliability

Research-grade pipeline for building a merged panel dataset to study heat-wave and wildfire-smoke interactions with U.S. electricity-grid reliability, including quality-assurance outputs.

PythonGeospatial dataPyTest