I am an AI/ML and statistical researcher developing reliable, interpretable computational methods for complex scientific problems—especially earthquakes, tectonics, geospatial risk and environmental hazards.
My work connects statistical inference, explainable machine learning, physics-based numerical modelling and geospatial data. I am particularly interested in models that do more than predict: they should quantify uncertainty, reflect domain knowledge and produce results that scientists and decision-makers can examine and trust.
- Physics-based modelling: used high-performance 3D numerical models to investigate how subduction-zone structure redistributes tectonic stress and influences large earthquakes.
- Statistical inference: applied Bayesian methods, maximum-likelihood estimation, probability distributions and uncertainty analysis to seismicity, plate motion and probabilistic hazard problems.
- Explainable machine learning: tested competing physical controls on megathrust seismicity, focusing on explanations that remain consistent with geophysical knowledge.
- Geospatial risk intelligence: combined GNSS-derived deformation, principal component analysis and population data to connect physical measurements with societal exposure.
- Trustworthy AI: design rigorous quantitative tasks, ground-truth solutions and evaluation criteria that expose reasoning failures in advanced language models.
I develop open, reproducible research software that turns methods from earthquake science into tools other researchers can inspect, validate and extend.
| Contribution | Reusable scientific capability |
|---|---|
| Automated GNSS strain-rate analysis | Automated processing of GNSS velocities into documented strain-rate estimates and geospatial outputs. |
Scientific issues, validation studies, documentation improvements and code contributions are welcome through each repository's GitHub workflow.
Explainable AI · supervised learning · dimensionality reduction · model evaluation · LLM post-training and benchmarking · human-in-the-loop validation
Bayesian inference · maximum-likelihood estimation · probabilistic modelling · uncertainty quantification · spatial statistics · reproducible analysis
Earthquake catalogues · GNSS and strain-rate analysis · tectonic modelling · geospatial analytics · numerical simulation · high-performance computing
Python · pandas · NumPy · scikit-learn · PyTorch · Rasterio · Folium · Streamlit · SQL · Docker · GitHub Actions
I hold a PhD in Geophysics from Monash University and completed postdoctoral research at Nanyang Technological University, Singapore. My research has progressed from physical models of subduction systems to statistical and explainable-AI analysis of earthquake processes, with publications in journals including Nature Communications, Journal of Geophysical Research: Solid Earth, Tectonics and Tectonophysics.
Across research and applied AI, the consistent goal is the same: turn complex scientific data into defensible evidence while being explicit about uncertainty and model limitations.
I am interested in applied research and technical collaborations involving:
- AI and machine learning for earthquakes, geohazards and Earth observation;
- Bayesian and statistical modelling of complex physical systems;
- explainable and uncertainty-aware scientific AI;
- geospatial risk, environmental intelligence and infrastructure resilience; and
- evaluation and assurance of AI systems used in technical domains.
If your work sits at the intersection of AI, statistics and natural hazards, feel free to connect through LinkedIn or email.