ESA Copilot – AI Assistant for Environmental Reports Tech: GPT-4o, Python, Node.js, React, RAG, Pinecone, Meilisearch, CV-based parsing

At Langan Engineering, I led the development of ESA Copilot, an internal AI system that accelerates Phase I Environmental Site Assessments (ESAs). The tool analyzes decades of environmental documents, including scanned PDFs, maps, regulatory records, and auto-generates report sections with audit-ready accuracy.

I built a custom RAG pipeline integrating LLMs, vector search, full-text search, and image-based document parsing. I also built the internal UI for prompt testing, report preview, and in-line editing.