Algorithms, AI, Statistics

Led the development of ML-powered mortgage servicing tools and hands off modeling capabilities. Book management, retention, customer segmentation, product offering, and data ingest.

• Led the development of various credit underwriting models for mid-market auto lenders. Responsibilities included data acquisition, target generation, model development, explainability analysis, and robustness testing. • Executed client-facing deliverables, including: guides for developing user intuition around our core technology, walkthroughs of experiments conducted by the modeling team, summary recommendations based on the modeling team’s findings, and model risk management (MRM) documentation.• Developed tools for automatically distilling explainability outputs into human-interpretable, actionable insights. Resulting tool reduced manual analysis by 95%.• Conducted research on the robustness of black box model explanations, providing theoretical and empirical guarantees on the reliability of the team's feature attribution methods. Results published in Zest AI's 'Robust Explainability in AI Models' white paper.

Researched novel techniques in machine learning explainability for diagnosing model failure and ensuring the semantic significance of decision making.

• Automated invoice generation to eliminate manual bookkeeping processes.• Analyzed market trends to generate potential verticals and client leads.• Composed email campaigns with an average 50 percent open, 5 percent response rate.

• Built a network of engaged data contributors to capture high-quality pricing data on consumer staples in multiple countries across Sub-Saharan Africa• Developed payment processes through SQL database methods to compensate data contributors in countries with complex capital regulations• Worked with mobile money organizations at the local level.• Provided analysis and feedback to data and sales teams within the company. This included: quality control analysis, consultation on client reporting, operations reviews, and growth capability estimations• Managed Premise's Instagram account (@premisedata) October '14 through January '15. Tasks included highlighting striking photos, analyzing hashtag relations and popularity in real time, and researching trending social and political topics to produce quality descriptions

• Helped grow, maintain, and manage teams of users in two cities in Nigeria (Lagos and Kano)• Provided analysis on growth tool platforms to better direct growth efforts• Studied data patterns as well as individual points to identify and eliminate user fraud and analyze usage patterns• Executed research to further growth operations in other GEOs

• Tracked and maintained inventory for lab networking devices • Studied and prepared proof of concept (POC) networking tests• Transitioned inventory management to the cloud to facilitate seamless usability• Prepared an educational course on electrostatic discharge for the systems testing team