I'm a Physics and Data Science student at UC San Diego researching scientific machine learning, physics-aware deep learning, and computational physics. My work spans PINN and neural-operator surrogates for thermal simulation, characteristic-aligned operators for transport-dominated PDEs, and MCMC-based analysis of lattice field theory — with published research in AI Interpretability and on Computational Physics. I'm looking for roles in SciML research, physics-informed ML, or quantitative research, where rigorous math meets real systems.
Bucket Labs
Prof. Tajana Rosing Lab / SEELab, UC San Diego
Prof. Javier Duarte Lab, UC San Diego
National Science Foundation REU Program
Simulated Sun-Earth-Moon three-body dynamics in pure Python with a symplectic leapfrog integrator that removed Euler's energy drift; implemented Metropolis-Hastings MCMC from scratch to recover an exoplanet's mass from noisy radial-velocity data; self-guided passion project.
Re-implemented PCA, k-means, EM, and regularized regression from scratch in NumPy with a FastAPI and React/TypeScript frontend; awarded 3rd Place at DinoCage, UCSD's Data Science Student Society hackathon.
Built WARP, the first model-agnostic adversarial robustness evaluation method for wildfire smoke detectors (YOLOv8n vs RT-DETR). Global Gaussian noise cut transformer mAP to 86.5% vs 50.8% for CNN; adversarial inputs reversed ~50% of smoke detections. Published in Fire.
University of California, San Diego