• Single-handedly designed and built an MLOps system on AWS for automated training, evaluation, and deployment of large-scale SageMaker ML models (with full end-to-end CI/CD) that ingest trillions of data points from Spark and produce billions of supply lead time predictions each day.• Performed a supply chain data study and retrained several XGBoost models with novel shipping data features, reducing predicted lead time by 20% (while maintaining accuracy) across 10,000,000+ products.• Rewrote a Python data processing module to use multi-threading (and other optimizations), reducing memory consumption by 95% and bringing our model evaluation time down from 4 hours to just 15 minutes.• Maintained and improved existing Java microservices/APIs, SQL queries, and Spark ETL pipelines.• Investigated and optimized my team's AWS storage and compute usage, saving over $100,000 in annual costs.• Performed various on-call duties like responding to tickets, fixing bugs, and monitoring alarms.

• Created a backtesting simulation in Python for quantitative trading strategies, along with a Django web interface for altering the strategies and viewing the resulting charts for P&L, drawdowns, and other risk/reward metrics.• Optimized existing quantitative trading strategies using Particle Swarm Optimization and Differential Evolution, improving their risk-adjusted returns by a factor of 2.• Created a butterfly spread trading strategy for agricultural commodity futures, resulting in an annual profit of $300,000 (500% return on 7% margin).• Developed a 10-day forecast of the lean hog cash price with scikit-learn, which was successfully used by traders to inform & improve their trading decisions.• Designed and trained a TensorFlow LSTM neural network to predict next-day prices for lean hog futures -- was successfully traded for 1 year with a directional accuracy of 60% and Sharpe Ratio of 2.5 (out-of-sample).• Single-handedly built and was responsible for the majority of the company's cloud infrastructure on AWS: EC2 production enviroments & webservers, RDS databases, and automated deployment pipelines.• Maintained and improved a Python task scheduling system which scrapes agricultural data reports in a variety of ways (regex, APIs), stores them in a database, and summarizes them in automated emails.• Built a variety of Django web dashboards for data visualization and scraping system control.• Trained, supervised, and mentored a Junior Software Developer.

• Contract to redesign client’s existing proprietary web app for their investment database• Discussed the specifications with the client and advised them on best way to improve the app• Revamped and added features to the existing CakePHP back-end• Managed the MySQL database and automated it using triggers• Rebuilt the front-end for a mobile-friendly responsive design using Bootstrap• Optimized the front-end and back-end for speed and usability