AI Integration Developer Needed for Existing MERN SaaS Platform

We have an existing *MERN-based SaaS platform* that is currently in use, and we are looking for an experienced *AI Integration Developer* to add intelligent features to the product. The core application, user management, database, and backend APIs are already developed. We do *not* need someone to rebuild the application from scratch. Your primary responsibility will be to understand the existing codebase and integrate AI capabilities into the current product in a clean and scalable way. We are looking for someone who can go beyond simply connecting an LLM API and can design reliable AI workflows that work with our existing application data, APIs, and business logic. ### What We Want to Build We want to introduce an AI assistant inside the platform that can: * Answer questions based on our internal documents and application data * Generate and summarize content using information available in the platform * Understand conversation context and maintain chat history * Retrieve relevant information before generating responses * Perform selected actions through our existing APIs * Trigger automated workflows based on user requests or events * Provide reliable responses with appropriate sources/references where applicable ### Your Responsibilities * Review the existing *React/Next.js + Node.js* codebase and understand the current architecture * Integrate *OpenAI, Claude or Gemini* APIs with our existing backend * Design and implement *RAG pipelines* for documents and internal knowledge * Set up embeddings, chunking, retrieval and a suitable *vector database* * Implement AI agents using *tool/function calling* * Connect agents with existing REST APIs so they can perform controlled actions * Implement conversation memory, context management and streaming responses * Integrate AI workflows with *n8n, webhooks and third-party services* * Improve prompts and retrieval strategies to achieve consistent responses * Implement basic guardrails, validation and fallback handling * Add logging and monitoring for AI requests, failures, latency and token usage * Optimize AI usage for *cost, speed and response quality* * Test AI functionality with different real-world scenarios before production deployment ### Technical Environment Our existing application uses: * *React / Next.js* * *Node.js / Express* * *MongoDB / PostgreSQL* * REST APIs * Docker-based deployment For the AI layer, we are open to recommendations regarding: * LLM provider * Vector database * RAG architecture * Agent framework * AI observability/evaluation tools ### Requirements We are looking for someone with proven hands-on experience in: * *LLM API integration* * *RAG and vector search* * Embeddings and semantic search * *AI agents and tool/function calling* * LangChain / LangGraph or equivalent * Node.js and REST API integrations * MongoDB/PostgreSQL * n8n or similar automation platforms * Docker and basic cloud deployment You should be comfortable working inside an *existing production codebase*, rather than building everything from scratch. ### Expected Deliverables 1. AI assistant integrated into the existing application 2. Production-ready RAG pipeline 3. AI agent with controlled API/tool access 4. Conversation history and context management 5. Required n8n/API automations 6. Error handling, guardrails and logging 7. AI usage/cost monitoring 8. Testing and documentation 9. Production deployment and handover We would prefer someone who can start with a small initial AI feature, demonstrate the approach, and then expand the integration across the platform if the implementation works well.