Lead expansion efforts of IBM's global hardware pricing ML application, which was responsible for a 30x reduction in deal cycle times, increased deal win-rates and improved gross profit margins. Devised and implemented new pricing approaches using gradient boosted quantile regression and probabilistic regression, which led to increased model performance and 20x volume reductions in ML models. Developed and refined hardware maintenance models that were responsible for reducing the need for the manual pricing of hardware maintenance service agreements. Contributed to BERT-based NLP model used for estimation of JIRA story points. Optimized deal similarity application that enabled business sales groups to make more informed pricing decisions via easy access to similar historical deals. Built out the model development portion of MLOps process (optimize source SQL queries, refactor online/offline code bases, update hyper-parameter tuning to use Optuna framework, develop standard model evaluation process across independent models).

Developed Publication/Subscription messaging service using Apache Kafka, Kubernetes, and Docker so that internal applications could communicate more effectively. Established data model for topics within Apache Kafka to ensure smooth flow of data from producer to consumer. Utilized Agile Development practices to facilitate development of Pub/Sub messaging service.

Put into production data engineering pipelines using SSIS and SQL Server to ensure bank analysts had necessary data for their reporting tasks. I performed data quality assurance tasks and validated data mappings to ensure senior data engineers could complete their data integration tasks.

Member of the City of San Diego's undergraduate data analytics program. I worked with fellow undergraduates to solve problems relating to the City of San Diego's repair-request application (Get It Done) to ensure better delivery of services citywide.