Financial Markets Developer in Austin, TX, United States

A system to ingest, clean, and process large data sets from industrial printers and predict when there will be a failure that would require service technicians to be deployed to reduce such calls.

Lead researcher and portfolio manager responsible for daily R&D and trading on a systematic EVA (economic value added)-based stock selection strategy. Managed one analyst. Implemented in both hedge fund and index and ETF formats. Created the Wilshire EVA 5000 Indexes.

Maintained TBs of data in a SQL database, which served a C++ trading algorithm to optimize the portfolio, a C# desktop app for data manipulation, Python, MATLAB, and R scripts for statistical analyses. Techniques utilized finance frameworks such as Fama-Macbeth regression, mean-variance optimization, multifactor risk modeling, and ML such as NLP and clustering.

Fuel Network Optimizer is a next-generation, AI-enabled software product for the oil and gas industry. It takes massive amounts of individual retail and wholesale fuel transactions as inputs, cleans them, and quickly computes a suite of business intelligence metrics (like volumes, profits, prices, etc.). It then combines these internal metrics with external factors such as weather, demographics, population density, traffic, proximity to major attractions, seasonality, and holiday effects, as well as features about each specific client retail site (like physical characteristics and customer loyalty), into machine learning models. The models predict sales, profits, etc. on a forward-looking basis, and also decompose history to increase understanding of what has transpired. Elaborate modeling of competitor behavior also exists, with a sophisticated price elasticity model that anticipates competitor reaction and recursively propagates price changes throughout the entire network of sites.