Senior military officer and operations research scientist applying high-performance computing and stochastic programming to the hardest resource allocation problems in defense.
I am a Lieutenant Colonel in the United States Marine Corps and a fellow in the Marine Corps Doctor of Philosophy Technical Program (PHDP-T) at the Naval Postgraduate School, competitively selected to conduct doctoral research in Operations Research. My work sits at the intersection of GPU computing, mathematical optimization, and defense decision-making.
My research investigates GPU-accelerated Splitting Conic Solver (SCS) as a competitive alternative to classical scenario decomposition for large-scale two-stage stochastic linear programs — with application to a novel stochastic programming formulation of Marine Corps Investment Portfolios under uncertainty. I have presented this work at the International Conference on Stochastic Programming (Paris, 2025), the SIAM Northern and Central California Sectional Conference (2025), and SIAM OP26 (Edinburgh, 2026).
Prior to my current assignment, I served as an Operations Research Analyst at Manpower and Reserve Affairs, HQMC, where I built analytical tools supporting the Marine Corps' $2.3 billion BAH appropriation program and applied optimization methods to officer talent management policy.
Dissertation and active research threads
Two-stage stochastic programming approach to portfolio optimization under uncertainty, applying GPU-accelerated SCS as a competitive alternative to classical scenario decomposition (Progressive Hedging). Applications include Marine Corps investment portfolios — a problem class currently addressed by discrete scenario selection rather than full stochastic programming.
Custom GPU kernels (CuPy RawKernel/RawModule) exploiting the block-angular structure of the two-stage stochastic program coefficient matrix — parallel sparse operations for K, Kᵀ, and (I + KᵀK) across the A, Tₛ, and Wₛ sub-matrices. Structure-aware implementation exploits block-angular sparsity for near-peak memory bandwidth. Presented at ICSP 2025 (Paris) and SIAM NorCal 2025.
Systematic evaluation of Ruiz normalization, adaptive/warped proximal operators, and Anderson acceleration on 2SSP instances (10–1,000 scenarios). Benchmarked against standard SCS on SIPLIB instances; combined enhancements reduce iterations by up to 67%. CG iteration limiting is the dominant practical accelerator.
Selected technical projects spanning GPU kernel optimization, operations research, and network optimization. Each project reflects hands-on work at the intersection of computing and applied mathematics.
GPU-accelerated computation of the Mandelbrot set in Julia using CUDA.jl. Implements FMA operations, loop unrolling (@unroll 128), early divergence exits, and memory-coalesced access patterns to approach the RTX 4080's peak memory bandwidth on this compute-intensive workload.
Min-cost flow model for crisis evacuation routing. Directed graph with 3 primary locations, 9 intermediate nodes, and 4 final destinations. Minimizes travel time under capacity constraints — applied to non-combatant evacuation operation (NEO) planning scenarios.
Statistical analysis of the Netflix content library using Principal Component Analysis and clustering algorithms. Explores content trends, genre distributions, and IMDb rating patterns across the streaming catalog using dimensionality reduction and unsupervised learning.
Replicated and extended three published cryptanalysis techniques for substitution cipher decryption. Combines neural sequence modeling with adaptive-mutation evolutionary search (DEAP). Achieved high decryption accuracy on test ciphertexts.
Open-source teaching resource for GPU-accelerated operations research methods. Worked examples in CuPy and CUDA covering sparse linear algebra primitives, custom kernels for structured LP and stochastic programming problems, and numerical methods for large-scale optimization. Designed for practitioners bridging OR methods and GPU computing.