About Me
I am a PhD Candidate in Physics with a focus on applied data science and machine learning. My work involves building and evaluating models on large-scale datasets, quantifying uncertainty, and translating complex quantitative problems into actionable insights. I am currently pursuing industry data science roles.
Physics Simulations
Stochastic 3D Multi-Field Systems
HPC / ETL
Distributed Pipelines
Deep Learning
ConvGRU Architectures
Analysis
Forecasting Turbulence
My work so far bridges the gap between theoretical physics and production-grade software engineering. In my PhD research, I build end-to-end ML pipelines for terabyte-scale 3D simulations, optimize code for HPC environments, and develop novel architectures to solve complex spatiotemporal problems. Whether it's forecasting molecular cloud collapse or optimizing distributed ETL pipelines, I thrive where data is massive, noisy, and physically constrained.
For my current work, I developed a multi-field fully 3D convGRU-UNet architecture to predict turbulent flows in star-forming regions. We improved predictive accuracy by 28% while enforcing statistical and physical constraints to ensure model validity. Learn more about the project here.
Fig 1. My Hybrid ConvGRU-UNet Architecture
My Timeline so far
Data Scientist, Graduate Research Fellow
Architecting 3D Spatiotemporal forecasting models (ConvGRU-UNet). Engineered distributed ETL pipelines for HDF5/Parquet, reducing training time from weeks to days via Multi-GPU DDP.
Applied Physics & ML Engineer
Designed Bayesian Neural Networks (BNN) for multi-target probabilistic regression, successfully modeling high-dimensional non-linear relationships with quantified uncertainty intervals
Lead Engineer (Computer Science Vertical)
Founded and led a cross-functional engineering team of 10+ to build a full-stack instrumentation facility. Developed Python-based automated signal processing pipelines for real-time spectral analysis and noise reduction.
Data Analyst
Investigated a statistically significant sample of Milky-Way/M31 analogs in TNG-5O, a set of large, cosmological magnetohydrodynamic galaxy simulations to quantify stellar Radial Migration.
Coordinator
Planned the logistics for meetings, conferences, seminars, lectures, and workshops. Served as the central point of contact and liaison between students, faculty and staff.
Secretary
Worked towards promoting space exploration through projects, conferences, and career development for students.
Scroll to explore history
Applied Data Science & Modeling Projects
Forecasting Complex System Evolution with Machine Learning
Built and evaluated deep learning models on large-scale simulation data to forecast the evolution of complex systems under uncertainty.
Project Importance: Scalable model evaluation, uncertainty-aware forecasting, and working with multi-terabyte, high-dimensional datasets.
Read more here
Panhandle Health Network: XGBoost based Appointment No-Show Predictor
End-to-end predictive system (XGBoost/Streamlit) for healthcare appointments that identifies high-risk patients and optimizes intervention thresholds to project ~$193K in revenue recovery.
Project Importance: Translating model performance into financial ROI, optimizing operational decision-making, and deploying interpretable AI tools for non-technical stakeholders.
Read more here
Probabilistic Regression and Uncertainty Quantification
Applied Bayesian machine learning methods to perform multi-target regression on noisy, high-dimensional data.
Project Importance: Uncertainty quantification, model interpretability, and statistical reasoning for decision-making under uncertainty.
Read more here
Statistical Analysis of Large-Scale Simulation Populations
Performed statistical analysis across hundreds of simulated systems to quantify population-level trends and variability.
Project Importance: Large-sample analysis, hypothesis testing, and extracting insights from noisy observational data.
Technical Arsenal
An interactive map of my production capabilities.