// people per project · agents on the tools
How the studio staffs your work
RandomSynergy is one operator plus a network, not a headcount. An engagement gets whichever mix the work needs: me end to end, specialist collaborators brought in per project, and an AI layer running on the studio’s own tools. You never pay for headcount a project doesn’t use.
There’s no team page here because there’s no fixed team, and being honest about that is the point. The receipt: 165+ repos on GitHub (RandomSynergy17), shipped this way.
The specialists
No fixed bench. Specialists come in project by project, matched to what a given build actually needs, then roll off when it ships. The network runs in both directions: I’m also the specialist other teams bring in, advising teams like Levels AV on AI integration and innovation, and running funding research and go-to-market for Abu Dhabi, UAE, and international programmes, serving Fortune 500 clients and small teams alike. Names are shared when they’re relevant to your engagement, not paraded here.
The AI layer
The studio runs on its own tools, and AI agents work through them the same way I do:
- GBrain: the studio’s self-hosted memory. Agents check it before answering and write back after working: 2,386 pages of decisions, infrastructure, and debugging history that survive the sessions that produced them.
- RNSNB: image generation as a service agents call directly, every image at a permanent URL.
- Publish and deploy pipelines: this site’s copy is drafted in markdown, run through automated review passes, and published by script, never by hand; code ships the same way.
Agents do the repeatable, well-specified work; I own the judgment calls. That division of labor is why a one-operator studio ships at this pace.
Tell me what you’re building, and I’ll tell you the mix it needs.