Hey, I've e built several production RAG and automation workflows in n8n, Supabase, and OpenAI. I can handle clean RSMeans ingestion, vector structuring, resume-safe batching, and a strict no-hallucination retrieval flow. I’ll deliver a reliable pipeline that maps SOW text to the correct RSMeans line items, grouped by CSI divisions, all within your budget
#Project: RSMeans → SOW Estimating Workflow
3 messages · Page 1 of 1 (latest)
Hi! I specialize in Structured RAG workflows using n8n and Supabase.
Your requirement for '1 row = 1 vector' is exactly the right architecture here, standard chunking would destroy the precision needed for cost items.
I can handle the 50k+ batch ingestion without timeouts and implement the Hybrid Search (Semantic + Keyword) needed to match your SOW to CSI divisions accurately.
My Portfolio: https://ai-labs-dev.github.io/Portfolio/
Ready to start immediately.
Hi, I am still available for the job.