- Identified missing fraud case data and collaborated with the vendor and internal data engineers to acquire and store case data- Optimized rules causing the large fraud ops case queue leading to a 33% case reduction and 73% increase in fraud decisions- Created and automated onboarding decision dashboards with KRIs and KPIs embedded for reporting and fraud investigations- Improved the onboarding acceptance rate from 65% to 80% by identifying false positive populations to safely approve- Analyzed and provided improvements to the document verification strategy capturing missing requirements and fraud risks

- Designed, developed and managed the implementation of Known User Score measuring similarity between user login attempts- Championed the Known User Score to senior management for implementation- Designed and lead experiment revealing vendor trustworthiness and cost reduction opportunities with improved performance- Lead the creation of an affiliation model leveraging internal data that captured 85% of vendor provided endpoint affiliation- Analyzed authentication events providing a deep understanding of user actions and identifying false positive decisions

- Performed analytics driving changes to client authentication processes reducing the Zelle fraud loss rate below peer average- Identified opportunities to improve ACH fraud recovery which could result in as much 80% reduction in ACH fraud loss - Classified general ledger losses to better understand the source of ACH and Wire fraud losses - Updated ACH and wire fraud detection rules to prevent evolving fraud schemes

- Utilized PCA and Decision Trees to create models that produced an ordered list of content for each customer- Performed Text Analytics on product descriptions extracting predictive information leveraged in our recommendation model- Developed automated audits to scan production tables for content that has missing or incorrect values - Built content ingestion jobs to help efficiently and accurately populate content before approving into production- Created models identifying Uverse customers with likelihood to move overriding the recommendation model results