-Foundation Project: Machine learning model developed in R to project NFL game outcomes and player statistics. Modeling techniques used include a random forest / multivariate logistic regression ensemble for primary projection as well as multivariate regression for other projections.-Monetization Channels: Produced written, audio, and video NFL content and analysis as a direct-to-consumer product and as a business-to-business product to two fantasy/gambling startups (BeerLife Sports, LineSniper).

-Transitioned the company’s Sentient Analysis model to a Large Language Model working with the RoBERTa model available from Hugging Face.-Leveraged an LLM to build a mode to interpret the context of ambiguous social media “reactions” (Facebook HAHA and WOW emojis) given the content of the “Parent” post; trained PyTorch model on Google CoLab GPU.-Created client dashboards in Tableau; highlight of this work was incorporating the White House’s Justice40 demographic maps to allow clients to target engagement in historically underserved communities.

-Developed an R Shiny App to project fuel savings from the use of Lightning’s braking system. Competing models were A/B tested, with the winning model built using h2o’s Random Forest in R.-Collaborated with the Engineering Team to design and execute Big Data studies on vehicle fleets.

-Used R and Tableau to conduct a study examining recruiting success based on demographic factors; uncovered notable insights on recruit productivity from particular states, specifically Illinois and Ohio.

-Built reports detailing audience metrics from a Druid database for internal customers (C-level executives).

-Developed data-driven opponent scouting reports based on data from film breakdowns.-Supported coaching staff in recruiting, practice and game day operations, and other day-to-day activities.

-Developed data-driven opponent scouting reports based on data from film breakdowns.-Supported coaching staff in recruiting, practice and game day operations, and other day-to-day activities.