About
I am a scientific software and machine learning engineer with a PhD in Physics from the University of Rhode Island (July 2026) and more than seven years of Python experience. My work sits at the intersection of numerical methods, research software engineering, and data-driven modeling.
During my doctoral research, I developed software for gravitational-wave science: reduced-order surrogate models for expensive numerical simulations, probabilistic models that carry waveform uncertainty into Bayesian inference, and distributed workflows for training and validation on CPU and GPU clusters. I worked with researchers in the LIGO Scientific Collaboration and the Simulating eXtreme Spacetimes Collaboration to turn scientific requirements into reusable Python APIs, tested packages, documentation, and shared analysis tools.
I also apply machine learning beyond physics. Recent work includes temporal graph neural networks for anti-money-laundering detection, a local-first semantic job-matching browser extension, and a deployed astronomy data API. Across these projects, I care about the same fundamentals: clear problem formulation, reproducible experiments, reliable software, and results that can be used by people other than the original developer.
I am currently pursuing research software engineering, scientific machine learning, and applied ML roles where I can build computational tools, improve research workflows, and work closely with domain experts.
What I Do
Scientific Software Engineering
I turn research methods into maintainable software. My projects use modular Python architectures, automated testing, continuous integration, package distribution, technical documentation, and tutorial notebooks so that models can move from an individual analysis into shared research workflows.
Numerical Modeling & Uncertainty Quantification
I develop reduced-order and probabilistic models for computationally expensive physical systems. My work combines singular value decomposition, empirical interpolation, Gaussian process regression, Bayesian inference, and careful numerical validation.
Performance Engineering & HPC
I profile scientific Python workflows to locate bottlenecks in data access, repeated model loading, and sequential computation. I use vectorization, caching, parallelism, and appropriate CPU or GPU resources to make research workloads more practical, then protect numerical behavior with regression tests. My HPC experience includes Slurm, MPI, parallel batch workflows, PyTorch, CuPy, Linux environments, and reproducible deployment with containers.
Applied Machine Learning
I build models and evaluation pipelines for graph data, temporal signals, anomaly detection, and severely imbalanced classification. I focus on metrics and operating points that reflect the real cost of false positives, missed detections, and limited review capacity.
Contact
I am open to research software engineering, scientific computing, and applied machine learning opportunities. If you are building computational tools for science or working on a data-intensive problem where numerical rigor and usable software both matter, I would be glad to talk.