n8n Automation Expert for AI Assistant

BEFORE READING : BDD IS IN FRENCH, UNDERSTANDING FRENCH IS PLUS.

I'm seeking an experienced n8n automation expert to help finalize a smart WhatsApp AI assistant. The core workflow is already in place, with essential components like conversation logging, user routing, and AI-powered responses functioning well. However, I need expert assistance to complete the vector-based memory system and finalize the embedding pipeline, focusing particularly on storing and retrieving embeddings.

The goal is to complete the project together, ideally through a live collaborative session (video call), or alternatively, to receive a clear explanation and documentation of what has been done once the integration is finalized. A smooth handoff and deep understanding of the final setup is essential.


What’s Already Done:

WhatsApp integration is functional (via webhook and middleware)
Full n8n workflow is already built (multi-module logic, GPT-based responses, logging, and memory logic)
Conversation logs are correctly stored in a Supabase database
Vector database (Pinecone) is connected and seeded with documents (needs verification)
Retrieval logic is partially implemented (RAG-style flow with vector search)
Supabase and Pinecone API keys, schema, and structure are already set up

Your Mission:

Audit the current memory setup (Supabase + Pinecone) and determine the best long-term option
(continue with Pinecone or move to Supabase PGVector if more stable or integrated)
Finalize the embedding pipeline inside n8n:
Add a chunking logic (LangChain node or custom JS logic)
Apply OpenAI Embedding to each chunk
Store vectors in Pinecone or Supabase
Finalize the retrieval flow:
On user input, embed the message
Perform similarity search in the vector DB
Inject results as context for OpenAI response
Review the conversation-saving logic to ensure consistency and memory reliability
(Optional) Advise on system optimization, scaling or long-term architecture

Required Skills:

Expertise with n8n workflow orchestration and advanced node logic
Solid experience with Supabase (table structure, auth, REST API, JSON handling)
Experience working with vector databases (Pinecone or PGVector)
Strong understanding of OpenAI Embedding models and prompt usage
Familiarity with LangChain integration in n8n (or able to replicate chunking logic)
Comfortable working with Function nodes, API calls, and data transformation
Bonus: understanding of RAG (Retrieval-Augmented Generation) principles