• Built and managed a team of 5-7 data scientists• Led the technical planning and architecting of Polaris and Rocket Frameworks for Trade Promotion Optimization (TPO) which is used by analysts to plan future promotions for product promotion groups (PPGs) such as 2pk of canned beans, buy-one-get-ones, gondola display, etc.• Managed ticket creation, assignment, hour estimation, effort tracking, spillover, and other metrics to optimize team health

• Created Loss of Revenue Models ($ per outage based on downtime and maintenance cost in human work force hours)• Led the creation of an AI Platform using the loss of revenue models to prototype the corporate ecosystem• Set up development and production environments using AWS, Databricks, Snowflake, JIRA, and Bitbucket

• Optimized queries for Out-of-Stock models for all items for all stores that Shipt delivers groceries to which covers 300+ metropolitan areas out of the 421 in the U.S. market• Created standards for new team members using ML Workflow Automation paradigms such as automated onboarding installs based on data science project, automated Docker environment images, and continuous integration continuous delivery (CICD) GitHub Workflows• Created automated metrics such as calculating out-of-stock rates and model F1 scores to monitor and manage system pricing in an automated model shadowing environment (promote and demote based on performance)

• Formulated Value-at-Risk (VaR) for 10M+ customers to create proprietary Exelon Utilities Risk Scores (EURS) which are used to forecast payment totals and bad debt reserves• Researched, diagramed, and initiated the construction of an end-to-end AI Platform to integrate existing and new technologies across the enterprise and streamline processes• Continued to build and support data science solutions while also advancing coworkers use cases when they were stuck or needed guidance• Configured LDAP and AirFlow for scheduling models in production• Led technical office hours and mentored new data scientists

• Developed a multi-turn closed domain conversational AI for the executive dashboard using a long short term memory (LSTM) neural network and the Rasa framework• Set the standard for coding best practices and principles involving customer-facing operations• Modeled customer electronic billing (eBill) and automatic payment (AutoPay) propensity scores• Created a recommendation system which pulls propensity scores from 30+ models across the enterprise to deliver targeted content to each customer• Created weekly unemployment rates on the county-level to account for pandemic pressures which current risk models at the time weren’t built to handle due to collection moratoriums• Aggregated energy usage to create load profiles for 10M+ residential and commercial customers for demand forecasting, customer segmentation, and anomaly detection• Built Augmented Risk Scoring (ARS) and Collection Cycle Effectiveness (CCE) models for evaluating customer financial risk which will save the company ~$8M+ per year as a baseline

• Developed novel 3D image processing algorithms to fully classify LiDAR point clouds using parallel programming, random forests, and custom ensembling• Built 3D models of electrical transmission circuits and their surrounding environments to optimize the use of company resources, time, and budget• Helped architect a customized VGG-16 neural network to classify 10 minute time series data from solar panels using TensorFlow to detect underperformance caused by hardware malfunctions and environmental factors• Created a custom model of site suitability to determine when to inactivate job sites for vegetation management based on meteorological data such as observed rainfall and evapotranspiration• Laid the foundation for a simulation of a multi-year vegetation management work order plan to solve the cost function for investing more money now to save money in the future based on circuit impact, risk, cost, and fines which is updated dynamically via a data feedback loop• Gave 40+ presentations which reached the Sr. Vice President level• Built a team of 15-20 people by winning an internal bid, led interviews for data science hires, and mentored new data scientists

• Collected key images from videos using Python/OpenCV• Successfully matched images using C# and Emgu CV, a .NET OpenCV wrapper• Built/trained a new model for processing in-game images using Tesseract-OCR

• Detected vehicles, pedestrians, road signs, road lines, and their distances using optical flow, histograms or oriented gradients (HOG), cascading object classifiers, Hough transforms, and stereo vision respectively• Participated in a U.S. Department of Energy (DOE) Advanced Vehicle Technology Competition (AVTC)• Designed image processing algorithms to improve vehicle efficiency and safety using C, C++, OpenCV, Matlab, and Simulink• Worked with 50+ students and professors from multiple disciplines on a 4-yr national competition presented by General Motors (GM) and the Argonne National Laboratory

• Supervised and mentored 40+ young adults in a military environment• Led team building exercises, morning devotionals, and managed the campgrounds, housing, laundry, kitchen, and dining facilities• Helped equip, inspire, and serve 750+ guests throughout the duration of their stay

• Built and configured a new computer server• Used C++/Python to aggregate and analyze healthcare data• 2013 Family Owned Small Business of the Year Award in Tennessee

• Assisted customers across the nation using call center software, scheduling software, and a medication therapy management database• Completed approx. 150-250 calls a day

• Organized an online marketing campaign involving social media• Advanced to the Sweet 16 in a university-wide contest to see which representative could enroll more prospective clients

• Learned various operational procedures of automobiles and the safety and efficiency of vehicle handling• Worked independently without manager oversight as well as in collaboration with other valets at times

• Maintained communication between approx. 4-5 workers in the kitchen and 8-10 workers on the floor• Learned point of sale systems