Data scientist with 8+ years of experience in tech and defense, and an M.S. in modeling and optimization. I build machine learning and GenAI systems that merge rigorous problem-solving with thoughtful design. My work spans deep learning, NLP, computer vision, and LLM applications—delivering production systems through cross-functional collaboration and strong product intuition.
Experience
Deloitte·Data Science Consultant | GenAI Engineer
Tempe, AZ · Sep 2023 – Present · 2.5 yrs
Trained and deployed a PyTorch neural network with Gaussian Mixture Models for benefits fraud detection—engineered features from demographic, geographic, and financial signals to flag anomalous claims; scaled to production on 32M+ records
Fine-tuned LLMs and built RAG pipelines with embeddings for enterprise GenAI apps—benefits claims assistant, IRS bill analyzer, liquidity platform—deployed end-to-end with Flask + React; serving thousands of daily users
Trained document classification models using GCP Document AI for IRS W-2 extraction; improved OCR accuracy on degraded scans through model tuning and preprocessing
Verra Mobility·Data Scientist | ML Engineer
Mesa, AZ · Feb 2019 – Aug 2022 · 3.5 yrs
Trained pricing optimization models using regression and simulation to balance customer adoption against financial exposure—drove $1.2M in annual revenue gains
Built gradient-boosted classifiers (XGBoost) to predict invoice default risk; enabled targeted collections interventions, reducing bad debt by 21%
Created revenue simulation for Ireland market entry using tourism and rental car patterns; presented findings to stakeholders, informed executive decision on international expansion
Raytheon·Optimization Systems Engineer | Data Scientist
Tucson, AZ · Mar 2015 – Jan 2019 · 4 yrs
Trained and deployed predictive models for real-time production monitoring—time series forecasting for lead times and anomaly detection for bottlenecks across assembly lines; adopted by 4+ factories, awarded "Project of the Year"
Built Monte Carlo simulation models in Python to digitize and optimize factory value-stream maps; enabled data-driven capacity planning, saving 500+ hours annually
Intel·Data Science Intern
Chandler, AZ · May 2013 – Jan 2014 · 8 mo
Built demand forecasting (random forest) and optimization (linear programming) models for capital equipment placement—delivered $1.8M/year in cost reductions
AI portrait platform using fine-tuned Stable Diffusion models. Users upload selfies and generate custom AI portraits in various styles. Full-stack application with Next.js frontend and Python backend.
Native macOS dictation app with real-time transcription. Optimized for speed and seamless OS integration—built to feel like part of macOS. Includes meeting recording with automatic transcription.
Find the best coffee shops for remote work and studying. Custom review analysis engine scores shops on two dimensions—coffee quality and work-friendliness—with 'coffee first' and 'work first' filter modes.
Market intelligence platform with interactive heatmaps, real-time market health indicators, stock screener, and aggregated financial news. Designed for quick daily market context.
Gmail inbox declutter tool. Automatically groups emails by sender, surfaces all your subscriptions in one place, and lets you mass unsubscribe with a single click.
Python · Gmail API · React · OAuth
AlphaSeek — FounderPrivate IP
ML-powered stock selection engine using LightGBM ensembles for alpha signal generation. Automated backtesting pipeline with momentum signals, risk management, and Schwab API integration.
Community platform for guitar tabs. Create, publish, and play along to tabs. Fork and improve others' work—GitHub meets Reddit for musicians. Upvote the best arrangements.
Research exploration into transformer architectures for financial time series. Custom hierarchical attention mechanism with local, intermediate, and global heads.