Resume

Summary

Recently returning to industry after my Ph.D.
Research in ML of physics, stochastics, and research engineering; experience with LLMs and RL
Lead maintainer of pysindy, a python ML package with 1.6k stars on github and hundreds of users globally
Prior professional experience leading data science and computer vision/listening projects for consulting clients.
Teamwork and leadership experience through seven years of Naval service in high-stakes environments.

Booz Allen Hamilton, Lead Scientist (Sep 2018—Sep 2021)
- Let a 5-engineer/statistician team to build time-series forecasting for mine-detection logistics, 5-50% planning accuracy improvement and secured multi-year contract renewal; designed experiments and automated statistical analysis, ran Agile, owned stakeholder and technical roadmap; stack: Python, scikit-learn, statsmodels
- Led a Machine Learning research effort to detect and track vessels from classified submarine acoustic recordings. Built unclassified effort, project Baya, to disseminate capability to company and published results in conference.
U.S. Navy — Lieutenant (May 2011 — November 2017)
- Lead operation planning for 10+ ship international exercises. Planned and oversaw humanitarian deployment Pacific Partnership 2015 and fleet exercises.
- Directed operations on Nimitz Strike Group in the Arabian Gulf, lead bridge team responsible for safe navigation in critical situations, and managed safe engineering plant operation across four deployments.

Ph.D., Applied Mathematics — University of Washington, 2025
- Led (& currently lead) pysindy, an ML library for learning dynamical systems with 1600+ stars on github and dozens of collaborating researchers and engineers throughout development lifecycle. Managed long-term engineering and scientific priorities and experiment design, reviewed PRs & implemented cutting edge methods around reinforcement learning, convex optimization, stochastic processes; Package used for studying metabolic processes, HVAC engineering, fusion reactor design, molecular angles, and more. stack: jax, scikit-learn, numpyro, gurobi, sympy, numpy, cvxpy and more.
- Reproduced LLM learning experiments from recent papers, implementing transformer networks in pytorch and evaluating model performance and training dynamics.
- Derived and demonstrated stochastic (Kalman-like) smoothing filter for Doppler sonar data from seagliders. Built and evaluated neural and statistical time-series models for reduced-order systems of fluid flows.
- Organized collaboration across several groups, including open source contributions of an experiment runner and production-quality features/bugfixes to numpy, matplotlib, pandas. Mentored peers and improved team dynamics to speed up and deliver research projects confidently.
B.S., Quantitative Economics/minor in Arabic, with honors — United States Naval Academy, 2011

Languages & Engineering: Python (production-quality), SQL, Research Engineering, Design Patterns, Agile Development
ML & AI: Research Planning, Experiment Design, PINNs, Neural ODEs, Large Language Models (LLM), Reinforcement Learning (RL), Time-Series Models, Bayesian Analysis
Frameworks & Tools: PyTorch, JAX, scikit-learn, NumPy, numpyro, Gurobi (optimization), cvx, statsmodels

A. Hsu, et. al; General Sparse Parametric System Identification via Collocation and Nonsmooth Optimization; In submission for Journal of Computational Physics
M. Peng, et. al; Local Stability Guarantees for Data-driven Quadratically Nonlinear Models; Accepted to Physics of Fluids, Oct 2025
I. Griss-Salas, et. al.; Coarse graining and reduced order models for plume dynamics; Accepted to Physics of Fluids, Oct 2025
J. Stevens-Haas, et. al.; Learning Nonlinear Dynamics Using Kalman Smoothing; IEEE Access, May 2024
J. Stevens-Haas, J. Keeton, R. Piazza. Detection and Classification of Subsurface Acoustic Targets using VGGish. Naval Applications of Machine Learning conference, Feb 2020.

Available upon request.