Aaron Glover

Engineering Manager, Machine Learning | Enterprise Products

May 2019 - Current

• Spearheaded the successful integration and doubling of two data science teams, scaling the team from 8 to 16.
• Engineered and deployed a predictive maintenance solution for critical industrial equipment, extending asset life and saving approximately $100,000 for every three months of extended use.
• Grew the team from inception to a team of 8 including data scientists, data engineers, PowerBI developers, and quantitative analysts.
• Architected and oversaw the production of over 50 predictive models for commodity trading, contributing to an 80% gain on our managed book in 2024.
• Directed the development of a 'no-code' model deployment system and a comprehensive model evaluation framework, significantly accelerating the path to production for new models.
• Championed the standardization and optimization of the DevOps lifecycle, implementing robust CI/CD pipelines, unit testing, and automated release schedules.

Machine Learning Engineer | Sanchez Energy

Apr 2017 - May 2019

• Developed a novel model fitting solution for determining a well's spontaneous (SP) log curve using peak detection methods and Kalman filters. The end result was used by the geophysicist team to identify potential oil field plays to target or acquire.
• Developed a Markov Chain Monte Carlo (MCMC) solution for simulating a well's decline curve and ultimate recovery. This augmented engineering's decision making on how much a well will produce over its lifetime and optimized field development and planning.
• Optimized an in house developed geophysics simulator in Python which decreased runtime by 6x and lines of code were reduced 10x.
• Contributed to the development of a multi model prediction framework for predicting well production. The solution was a significant improvement on the industry standard decline curve fitting.
• Implemented a real time alert for detecting tubing leaks which resulted in a cost savings in the six figures. An industry standard deterministic model was required by Engineering, the model was optimized by sampling the search space with a Tree-structured Parzen Estimator.

Data Analytics Engineer | Occidental Petroleum

Jan 2012 - Mar 2017

• Developed a Monte Carlo simulation in python to determine the optimal number of workover rigs for a given field. This solution was implemented in multiple fields across Texas and California with a savings in the high six figures.
• Developed over 350 SSRS reports and Spotfire dashboards for every department in operations, engineering, and production accounting.
• Designed and developed multiple SSAS cubes for operational and well servicing data, query times were reduced 1,000x.
• Maintained and enhanced the main Operational Data Store (ODS) used company wide for production reporting.
• Automated the delivery of partner reports utilizing SSRS and SQL which resulted in an 80% reduction in man hours.

M.S. in Analytics

Texas A&M, May 2017

GPA: 3.8

Thesis: Predicting the likelihood of ESP well failures utilizing survival analysis and gradient boosting.