Nathan Nelson

I've built AI-assisted media intelligence tools at Relo Metrics, helping users search sponsorship, campaign, audience, and media information through natural-language questions. I developed Python retrieval workflows for document ingestion, chunking, embeddings, metadata filtering, vector search, reranking, and LLM responses.

I built FastAPI services and LangChain and LangGraph conversational workflows, while validating document extraction and summarization output before it reached other application components. I also used CloudWatch, Prometheus, and Grafana to investigate slow responses, API failures, retrieval issues, and model behavior after release.

At ZeOmega, I developed Python and Java backend services for healthcare data and knowledge applications, including REST APIs, document processing, data access, and integrations. I later worked on NLP features for healthcare documents, data-processing jobs, query tuning, Redis caching, Docker, AWS, and Kubernetes deployments.

Earlier, I built consumer, product, and operational data applications at Colgate-Palmolive and internal operations services at Amazon. Across my work, I take systems from early prototypes to production alongside product, data, security, and engineering teams.