AI Ticketing Platform (as a Service)

Overview:

A real-time, multi-tenant AI-assisted support ticketing platform, built on the same architectural foundation as Octavia AI CMS, supporting 7 languages with full RTL/LTR support.

My Role:

Co-founder and full-stack developer at OctaviaTech, responsible for extending the shared multi-tenant backend to support real-time updates, and for implementing live synchronization across the frontend.

Technologies:

Frontend:

•       React

•       Tailwind CSS

•       shadcn/ui

•       Motion

•       Recharts

•       TypeScript

•       Socket.IO

•       Docker

Backend:

•       Dedicated AI service

•       Express

•       Node.js

•       Kafka

•       Mongoose

•       Pino

•       TypeScript

•       Redis

•       RBAC

•       Socket.IO

•       Docker

Challenges & Achievements:

•       Implemented real-time ticket updates via Socket.IO across both frontend and backend, so every open session reflects changes immediately rather than on the next refresh.

•       Designed the underlying state model so two agents can act on the same ticket concurrently without one silently overwriting the other.

•       Used Kafka to decouple AI-driven ticket triage and categorization from the parts of the system that need to feel instant, such as new replies appearing on screen.

•       Used Redis to back fast, cross-instance state including session presence and live ticket counts.

•       Enforced tenant-scoped RBAC across every socket event, cached value, and AI-generated suggestion, with zero cross-tenant exceptions.

Project link: Internal product — not yet publicly listed.