Sohini Dasgupta

I am a Ph.D. candidate focused on computational geophysics and machine learning, where I develop physics-driven carbon monitoring methods and surrogate models for carbon sequestration and storage. My interests include scientific computing, numerical algorithms, and rock physics, and I actively apply machine learning to tackle large-scale joint inversion problems. I am also keen on statistical analysis (multi-modal) for reservoir characterization, exploration, and energy transition challenges. Prior to my doctoral studies, I worked as an exploration geophysicist, where I acquired valuable skills in integrated seismic interpretation, seismic attribute analysis, velocity modeling, and well planning and execution.

Current Research Projects

Rock physics constrained machine learning models for carbon monitoring

Physics guided machine learning workflows for joint inversion

Past Research Projects

Statistical analysis on multi-variate data from a generative AI seismic inversion for reservoir characterization

Diffusion learning in seismic inversion

Prediction of CO2 saturation and elastic properties from capillary pressure-based rock physics model

Developing a 2D map-based inversion workflow using physics-informed machine learning method

Prediction of minimum mud weight for safe well drilling and sensitivity analysis for tectonic stress in NE part of India