Senior Principal Machine Learning Engineer

I serve as Team Champion for a team of 15 machine learning, software, and data engineers. I own the technical direction for engineering initiatives, including prioritization, architecture, system design, design reviews, code reviews, reliability standards, scalability, and MLOps practices.

I lead development of Rocket Logic, an AI platform that automates the homebuying process. The platform classifies 70% of 1.5 million monthly documents, reduces loan processing time by 25%, saves more than 9,000 manual hours each month, and supports loan closings 2.5 times faster than the industry average. I also built an LLM-agent platform that delivers real-time, dynamically generated analyses, data summaries, insights, and visualizations.

I architected a Market Impact model stack that adjusts primary hedge models for market risk between rate lock and loan close. I also built an enterprise Feature Store supporting training, scoring, and reporting across Rocket Exchange through event-driven and scheduled workflows, dynamically partitioned processing, and data-governance controls.

I co-authored Rocket's multi-year data and AI strategy, led AI design reviews, and provided architectural guidance across the organization. I led the migration of 400 data scientists, analysts, and engineers from local environments to an AWS-based platform, founded the RKT Machine Learning Community, and regularly presented on AI and ML to leadership and internal technology forums.

I led development of a geospatial machine learning service that used nationwide location data to identify homes near solar installations. The service powered solar loan campaigns and identified high-propensity purchase and refinance leads for a new business line.

I designed and delivered a real-time, event-driven, ML-library-agnostic hedge model scoring service on AWS. The service scored approximately 1.5 million records through 34 dynamically chained models in six seconds, reducing processing time from 12 minutes and cutting error rates by 50%. It also reduced single-model deployment time from two business days to five minutes and created approximately $10 million in value.

I designed a real-time, asynchronous, microservice-based machine learning prediction engine capable of concurrent multi-model scoring. I built Auto EPI, an AutoML infrastructure that automated data sourcing, feature selection, algorithm selection, hyperparameter tuning, deployment, monitoring, alerting, and model-drift detection.

The Auto EPI platform monitored models across more than $50 billion in daily volume. I also developed an NLP-based candidate-ranking system for recruiting, led ML and big-data engineers, delivered ML solutions for four business teams, mentored engineers and data scientists, and contributed to hiring.

I built an in-house AutoML platform using Spark ML to process petabytes of data for multi-billion-dollar hedging operations. The platform used ensemble methods, support vector machines, neural networks, anomaly detection, and dimensionality-reduction techniques to mitigate risk.

I devised continuous automated statistical testing to monitor and improve production model reliability. I also published pip-installable data science packages that increased algorithm reuse and team productivity.

For an NSF-funded rapid blood typing microdevice project, I developed SVM, ANN, CNN, and RNN models to predict ABO-Rh blood types. I created real-time R Shiny dashboards and deployed the model using a Flask and OpenAPI REST API served through Gunicorn.

For the Great Lakes Research Center's semantic sentence equivalence research, I built an NLTK preprocessing pipeline using word2vec embeddings and a Manhattan-distance LSTM model. The solution achieved 91% accuracy.

I drove more than $10 million in monthly growth in foreign deposits by building third-party integrations with VISA, Mastercard, Western Union, WorldRemit, and MoneyGram. These integrations expanded banking and remittance capabilities for customers.

I developed Django REST APIs supporting internet banking, Android and iOS mobile banking, and USSD banking for more than five million customers worldwide. I also managed the bank's internet and mobile banking platforms.

Graduated with a 4.0/4.0 GPA and was recognized as the top graduate student in Data Science. Recipient of the Dave House Graduate Fellowship from 2016 to 2018.