I'm a PhD candidate at Georgia Tech working at the intersection of scientific machine learning, multi-physics simulation, and digital twin systems — with applications in energy, geoscience, and large-scale physical systems.
- Inverse Problems & Bayesian Inference — amortized and sequential posterior estimation, UQ for high-dimensional inversion
- Multi-Physics Simulation — two-phase Darcy flow, wave propagation, and coupled flow-imaging workflows
- Digital Twin Systems — physics-based simulation coupled with generative models for real-time forecasting and decision support
- Scientific ML & Surrogate Modeling — Fourier Neural Operators, conditional diffusion, and transformers for coupled PDE systems
- Multimodal Learning for Physical Systems — late-fusion of spatial imaging, time-series sensor data, and physical priors for downstream inference and decision support
- HPC & Scientific Computing — large-scale Julia / Python pipelines on GPU and distributed systems
-
Twin4GCS.jl Digital twin framework for underground energy storage. Couples generative posterior models with multi-physics simulation (two-phase Darcy flow + wave-equation imaging) for sequential Bayesian inference of subsurface saturation and permeability fields. (Julia)
-
neural-flow-surrogates Benchmark comparing Fourier Neural Operators, conditional diffusion U-Nets, and Diffusion Transformers as autoregressive surrogates for coupled multiphase porous-media flow on 256×512 grids. Reports per-channel error and calibrated uncertainty against a JutulDarcy reference simulator. (Python / PyTorch)
-
Nonlinear-JRM Time-lapse seismic imaging via the Joint Recovery Method with a fully nonlinear wave-equation forward operator — consistent multi-vintage reconstruction from sparse, noisy observations. (Julia)
-
ccs8803-env_JUDI & ccs8803-env_Jutul Containerized compute environments (JUDI + Jutul + SLIM stack) used in the CCS 8803 graduate course at Georgia Tech for hands-on seismic modeling, imaging, and reservoir simulation exercises. (Docker / Julia)
-
💼 LinkedIn · 🐙 GitHub · 🎓 Google Scholar