Bridging deep learning with the Earth
I develop Scientific Machine Learning methods for seismic monitoring, source localization, and subsurface characterization — connecting neural operators, generative models, and physics-informed learning with real geophysical applications.
Research interests
My work sits at the intersection of Scientific Machine Learning and computational geophysics, bridging theoretical deep learning with robust physical applications.
Neural operators
Fourier Neural Operators (FNO), Graph Neural Operators (GNO), and foundational models for scientific data — pushing resolution-aware, physics-consistent learning.
Inverse problems
Diffusion models (DDPM, I2SB) for joint parameter estimation, velocity–source inversion, and uncertainty quantification in ill-posed geophysical problems.
Edge-deployable AI
Lightweight neural operators for real-time microseismic monitoring, DAS data compression, and automated event detection at the edge.
Applications
Sustainable geo-energy (geothermal, CO₂ sequestration), continuous subsurface monitoring, and data assimilation for digital twins.
Selected publications
Recent peer-reviewed work spanning neural operators, explainable AI, and microseismic monitoring. Browse all publications.
Explainable AI for microseismic event detection
Grad-CAM and SHAP-based tools for interpreting PhaseNet decisions, enabling transparent and robust microseismic event detection under noise.
Explainable AIEvent detectionAI in Geosciences, 2026
Microseismic source localization using FNO
Fourier Neural Operator for real-time microseismic source localization with application to field data from the Utah FORGE geothermal site.
Neural operatorsLocalizationIEEE TGRS, 2025
Deep generative I2SB for joint inversion
A Diffusion-model framework for joint microseismic source localization and velocity inversion with uncertainty quantification for hydraulic fracturing monitoring.
Diffusion modelsInverse problemsUnder review, 2026
Seismic event classification with lightweight FNO
Parameter-efficient Fourier Neural Operator model for seismic event classification, validated across out-of-distribution benchmark datasets.
FNOClassificationGeophysical Prospecting, 2026
7+ Publications
5 years Industry (TATNEFT)
3 degrees BSc · MSc · PhD (2026)
3 universities KFU · CSU · KFUPM
Let’s collaborate
I am interested in research collaborations in Scientific Machine Learning, neural operators, generative models, and their geophysical applications. Feel free to reach out.