Research
Computational geomechanics, numerical modelling, and AI-enabled engineering workflows — published work, prototypes, and in-progress thinking.
I gave five frontier AI models the same one-sentence prompt: build a website explaining coupled thermo-hydro-mechanical processes in porous media, with beautiful and interactive illustrations. The results ranged from genuinely impressive to broken. All five are...
Exploring the capabilities of the newly released Gemini 3.1 Pro by generating a complex, interactive 3D visualization of soil constitutive models using Three.js and Plotly.
A reflection on the rapid evolution of AI tool development—from manual function calling to the Model Context Protocol (MCP) and Agent Skills—and why staying ahead means building for adaptability, not just the latest standard.
Are digital twins genuinely intelligent systems for geohazard management, or are they just glorified 3D models? Exploring the challenges and proposing a hybrid framework.
Introducing GeoSim.AI - a suite of AI assistants for numerical simulations in geomechanics that demonstrates the transformative potential of generative AI in geotechnical engineering.
Physics-informed neural networks (PINNs) for forward and inverse problems in one-dimensional consolidation are presented. The governing partial differential equation is applied as a constraint in the neural network using automatic differentiation. The model performance is demonstrated...
The fundamental concepts and formulations in porous media theory are presented, covering volume fraction concept, conservation laws, thermodynamics, phase transitions, and coupled problem formulations for hydraulic, hydro-mechanical, and thermo-hydro-mechanical processes.