Stephen Talley

I’m an AI engineer with a data science foundation building productionAI applications, LLM evaluation pipelines, and Python backend services in healthcare and pharmaceutical settings. I bring AI workflows from prototype to production by pairing strong backend engineering with rigorous model evaluation.

At Idea Evolver, I built Django back-end services and RESTAPIs to deliver AI-driven endpoints to front-end applications. I operationalized a proof-of-concept into robust, scalable, production-ready ACNU feasibility scorecards, and I architected and deployed a production AI visibility / LLM evaluation pipeline processing up to 2,500 brand- or condition-specific queries per project using an LLM-as-a-judge framework.

I also fine-tuned OpenAI and Gemini LLMs as supervised multi-label classifiers for FDA Orange and Purple Book pharmaceutical data, mapping medications to 30+ therapeutic areas with train/validation/test splits. Previously, at ConcertAI, I led data quality control reaching 98%+ accuracy and 20% runtime savings, and I owned reporting, QC, and delivery for a $200K+ pharmaceutical account; earlier at UNC Wilmington I built Python/R genome-wide association tools and processed 2+ TB of genomic data on the Stampede supercomputer.