Backend Developer Needed for Real-Time AI Chat App (Supabase + OpenAI + RAG)

We’re building a custom AI chat assistant inside our app using Svelte (frontend) and Supabase (backend).

We're looking for a skilled backend developer to help us build the core infrastructure for:

Real-time messaging using Upgrade to see actual info (or SSE)

AI responses via OpenAI (GPT-4-turbo, streamed)

Document-based contextual replies using Supabase Vector DB (RAG pattern)

This is a short-term fixed-price project, with opportunity for ongoing work if it’s a good fit.

What You’ll Be Doing:
Set up a real-time backend server using Node.js + Upgrade to see actual info (or SSE)

Integrate OpenAI API with streamed responses to a client

Build RAG system using Supabase Vector DB:

Chunk documents

Create embeddings

Search based on user input

Create endpoints for:

Chat history

Recent sessions

Saving/retrieving user messages

Deploy backend to a platform like Upgrade to see actual info, Render, or Railway

Deliver with clean, modular code and a short handoff doc

Ideal Experience:
Node.js (Express or Fastify)

Upgrade to see actual info or real-time web backends

Supabase (especially Vector DB & Postgres)

OpenAI GPT-4-turbo streaming API

RAG implementation with real documents

API development and deployment

Deliverables:
Real-time AI chat backend (ready for frontend integration)

RAG-powered response engine

Session-aware chat storage & retrieval APIs

Deployable codebase with docs

Timeline:
Start immediately. Completion expected in 10–15 working days.