RAG must be developed as an independent regulatory validation engine running after FINAL MERGE, using a closed-domain approach that operates only on uploaded official documents without external web search. It should run after final_merged_text is completed and Vision results are appended, connected from n8n only via a Side-Car API call. RAG must be deployed as a separate Docker container with a vector database in channel-specific namespaces already made in current workflow Input data should include final_merged_text and Vision tags, and RAG must not influence generation logic, only validate final outputs. The output must be a structured JSON validation report containing legal references, not just OK/NG. Because this is a closed-document RAG structure, it provides high accuracy and relatively low development complexity.! RAG Engine Construction & Data Training Integration We have an existing n8n-based AI video automation system. The task is to develop the features listed below and ensure seamless integration with the current system. UI designs provided. Difficulty: Low Scope: Review the existing design of the Gosed RAG engine for large document training. Modify and connect data pipelines to ensure seamless integration with downstream n8n workflows and UI/UX connections. [Mandatory Deliverable]: A Google Sheets-based manual including step-by-step screenshots, prompts, and configuration values (Video + Text). [Mandatory] tell me your portfolio related to this task. and Tell me price and timeline.