Production AI support-agent SDK for an enterprise client

Built and operate a TypeScript AI agent SDK that connects an enterprise chat platform to a ticketing system with a RAG knowledge base, handling real support traffic in production.

• Node.js
• TypeScript
• RAG
• LLM routing
• systemd

Context

An enterprise client needed their support chat channel connected to their internal ticketing system, with an AI layer able to answer from a knowledge base instead of only routing tickets.

Challenge

Build something that runs reliably as production infrastructure — not a demo — including session resilience across restarts and multi-tenant credential routing, since the same SDK serves more than one internal bot.

What I did

• Designed and built a Node.js/TypeScript SDK bridging the chat platform, the ticketing system, and a RAG knowledge base.
• Added an LLM-based intent classifier as a routing fallback for messages that don’t match a known flow.
• Deployed it as a systemd-managed service and handled real operational issues (identity-sync failures, session recovery) as they came up in production.

Stack

Node.js, TypeScript, RAG pipeline, LLM-based intent classification, systemd.

Outcome

A live production system handling real support conversations, with multi-tenant routing so it can serve more than one bot/client from the same codebase.