Selected work / AI enablement
Content operations, designed as a real system.
An independent platform I designed, deployed and operated to connect ideas, assets, AI-assisted production, scheduling and reporting in one modular workflow.
Independent platform · operated end-to-end for a paying clientThe problem
Content work often loses momentum between the first idea and the published asset. Files sit in the wrong place, editorial context gets detached from the media, approvals disappear into messages and reporting becomes an afterthought.
What I owned
I owned the server setup, Linux and Docker deployment, workflow architecture and production operation. The platform combined n8n, OpenAI/ChatGPT, Gemini, Notion, Google Drive and Postiz, with Notion as the operating source of truth and Drive as the asset store.
What the system connects
The architecture covers idea capture, transcription and subtitle handling, title and description generation, structured prompts, asset classification, archiving, production tracking, scheduling, publishing and consolidated reporting. Each component was designed to hand useful context to the next rather than act as a disconnected automation.
Technical scope
- 14workflow exports in the documented architecture
- 287workflow nodes across the exports
- 288directed connections between nodes
- 1reusable operating model from input to reporting
Why it matters
This project shows the kind of AI enablement I care about: not using a model as a novelty, but building the operating environment, controls and handoffs that let people use AI as part of reliable creative work.