cd ../work
StreamBox Support
An intelligent technical support platform with a multi-agent AI workflow
- Role
- Full-Stack Developer Intern at ZetaBox
- Period
- Jan 2026 – May 2026
// overview
What it is
A web-based technical support platform for ticket management, project tracking, dashboards and automated reporting. Alongside the front end and back-end features, I implemented key components of its multi-agent AI workflow.
What I built
- Built responsive interfaces with React, TypeScript, Vite, Tailwind CSS, React Query, Zustand and Recharts.
- Integrated secure back-end features across two services using Strapi 5, Fastify, RESTful APIs, JWT authentication, RBAC, PostgreSQL and Redis/BullMQ.
- Implemented key components of the multi-agent AI workflow, from ticket triage to GitHub pull request generation.
- Supported containerized deployment using Docker, CI/CD pipelines, Dokploy, an OVH VPS and DNS configuration.
Stack
- React
- TypeScript
- Vite
- Tailwind CSS
- React Query
- Zustand
- Recharts
- Strapi 5
- Fastify
- PostgreSQL
- Redis/BullMQ
- Docker
- Dokploy
// architecture
How it fits together
A layered overview of the parts named on my CV. Select a part to read what it does.
Front end
Two back-end services
Data
Deployment
Front end
React + TypeScript
Responsive interfaces, built with Vite and Tailwind CSS.
// workflow
The multi-agent workflow
Ticket triage > Code analysis > Task planning > Human validation > Execution tracking > GitHub pull request
// verification
Automated test results
- 30/30
- Front-end tests
- 30/30
- API and security tests
- 14/14
- E2E, accessibility and browser tests
- 4/4
- Smoke tests