AI Researcher & Engineer

Built and owned, end to end, a conversational AI assistant embedded in the team's Basketball Operations infrastructure, giving analysts and decision-makers natural-language access to internal data and tools instead of writing queries themselves.

• •Designed every layer: an async Django Channels ASGI agent runtime streaming responses over HTTP, a hybrid dense and sparse retrieval engine (BGE-M3 and Qdrant), and a permission-aware tool layer enforced in the backend so the model never touches data a user isn't cleared to see
• •Built the durable session persistence layer on PostgreSQL and the analyst-facing frontend
• •Reported to the Director of Data Solutions, Insight & Foresight

Helped build GeoSight, an AI-powered digital public infrastructure dashboard for real-time SDG monitoring, unifying international indicator feeds into interactive geospatial analytics.

• •Orchestrated agentic workflows over the Google and UN Data Commons APIs and UNICEF SDMX data streams to turn natural-language questions into interactive dashboards
• •Presented the platform to UN and Google Data Commons stakeholders as part of a UNICEF-featured builder team

AI/ML Researcher

Dakota State University

Researcher in Prof. Andy Behrens's lab, building explainable AI pipelines, clinical decision-support middleware, and institutional automation systems.

• •EXAIM: built a transparency middleware for real-time multi-agent clinical decision support, validated on 40 clinical cases; accepted as first-author, peer-reviewed research to IEEE ICHI 2026
• •RegFlow: built a three-tier extraction pipeline that automated the DSU Registrar's graduation-audit process, FERPA-compliant by construction and in active institutional production
• •EXAI-KG: built a containerized pipeline that populated over 680,000 clinical records into the OMOP CDM standard, supporting explainable AI research on a synthetic patient cohort