Aaron Casey

NLP Engineer with a passion for building systems that extract meaning from messy, real-world text. I specialize in clinical NLP, designing LLM-based pipelines that turn unstructured medical documents into structured, actionable data. I'm driven by work that reduces suffering at scale, and I'm looking for mission-aligned roles in patient care, global health, or computational biology where my expertise in language understanding and knowledge representation can make a tangible difference.

I'm currently an NLP Engineer IV at Availity, where I build clinical NLP systems that automate prior authorization workflows in healthcare. My day-to-day involves engineering LLM-based extraction pipelines that identify medical concepts in clinical text, normalize them against curated UMLS ontologies, and output structured JSON for downstream clinical reasoning. I've worked across the full stack of production NLP — from transformer fine-tuning and GRPO optimization to knowledge graph construction, inference optimization with vLLM, and deploying models at scale. My recent focus has been on clinical logic generation, pushing the boundary on how accurately and efficiently LLMs can reason over healthcare data.

Before transitioning into engineering, I spent three years working as a medical assistant, which gave me firsthand experience with clinical workflows, patient documentation, and the pain points that healthcare professionals face daily. That background informs how I design NLP systems. I'm not just processing medical text in the abstract, I've lived the workflows these tools are meant to improve.

M.S. in Computer Science from North Carolina State University, with a foundation in machine learning, NLP, and systems design. B.S. in Biology from UNC Asheville, which gives me a strong grounding in the life sciences, particularly valuable when working at the intersection of language AI and healthcare or computational biology.