I’m Akshay, a researcher based in Germany using Deep Learning to bridge the gap between geophysics and geology! I have a deep interest in geophysical data integration and application of various geophysical methods to different problems in geophysics and other overlapping areas. I like working on problems of all scales from mineral exploration to plate tectonics, and using geophysics combined with geology to improve our understanding of our planet. Feel free to reach out!
- Potential field geophysics and structural geology
- Neural fields for geology and geophysics
- Spatial neural networks for physics informed interpolation
- Geophysical forward models and inversions
Neural Fourier–feature fields for scalar potentials and Hessian-based tensors (e.g., FTG). Includes:
- multi-scale RFF encoders with optional harmonic decay
- clean, well-documented training loop with Poisson-disk regularization
- first/second-order autodiff utilities (gradient & Hessian)
- a small ensemble wrapper and convenient plotting helpers