Senior Data Scientist
CurrentApplied AI- Built ML-based customer-facing products for spend management software to improve user experience and increase compliance.- Developed internal tooling to evaluate the performance of LLMs and ReAct agents. Tooling used by AI org to increase developer productivity, improve model performance, and prevent regressions.Fincrime- Worked with cross functional partners to determine effective fraud prevention policies while enabling growth.- Developed machine learning models to improve and automate business decisions relating to fraud risk.- Responsible for entire model development process including data set generation, exploratory analysis, feature engineering, model training, evaluation, deployment, monitoring and alerting, and analysis of business impact. - Supported engineering pipeline which pulled, stored, transformed, and fed data into model in production. - Work demonstrated clear impact: after deploying new version of model and adjusting related policies, fraud losses dropped by around $1 million per month.