Hi everyone 
,
We’re working on a pipeline using n8n + Supabase to process safety non-conformities (loaded from Google Sheets). The goal is for our AI agent to infer risk levels and suggest corrective actions based on the incident descriptions.
However, we’re facing these issues:
• The RAG loses context and gives incoherent outputs.
• The prompting is not structured enough (we need risk level + justification + action).
• The overall pipeline isn’t stable, and our internal devs are stuck (some work with code, others with no-code in n8n).
We need:
• Fast consulting (1–2 weeks max).
• Someone with real-world experience in production RAG pipelines, embeddings, vector databases, and LLM prompting.
• We’re open to hourly or per-deliverable contracts – we need to start ASAP (we have 2 clients waiting).
Please DM me if you’re interested or drop your contact info. Thanks!