AI Operations Lead

We built an AI Chief of Staff out of Claude, ran a real Australian B2B company on it for six months, and now we sell it. 48 custom skills. 18 operating rules. 4 custom integrations. 40+ always-on workers. 240+ memory files. Runs at roughly USD $90 per month in AI credits. It runs the daily operating jobs across sales, marketing, customer intelligence, finance, meeting intelligence, and board reporting. Now we are packaging it as Seby and deploying it for other companies. Our first client is a specialist recruitment firm. There is a queue behind them.

You will be the person who runs this. Not the person who codes it: the person who deploys it, judges it, iterates it, and trains clients on it. You will talk to Claude all day, decide whether its output is done, draft, or wrong, and push back when its plan is off.

This is not a coding role. It is not a support role. It is not a junior position. It is the role of an operator who understands how an AI-native business runs end-to-end.

Why this role is rare

You will exit this role as one of a handful of people in the global market who can read a business, map it to an AI operating system, and deploy that system into it. That capability does not exist in a course. It does not exist in a degree. It exists in roughly twenty people on the planet today, and almost none of them got there by accident. They got there by running this exact system on a real company. You will be the second person to do that.

You will report directly to Michael, the CEO. No middle layer. Daily contact. Full context.

By month six, you will have deployed the system into at least two paying client businesses and will be running the day-to-day operating rhythm inside Seby itself. By month twelve, you will be running deployments without Michael in the room.

What you'll own

In outcomes, not activities.

By end of month 1

You understand the system end-to-end. You have read every skill, every rule, every handover document. You have run the daily operating rhythm yourself for two consecutive weeks. You can explain any part of the system in plain English to a non-technical founder.

By end of month 3

You have led one full client deployment. Taken the system, adapted it to a new business's workflow, and handed it over working. The client is operating it themselves with weekly check-ins from you.

By end of month 6

You are running the Seby operating rhythm internally: inbox triage, customer intelligence, meeting debriefs, scheduled-task health, pipeline reviews. You have led two more client deployments. You are independently making architectural decisions about which client problems get automated next.

What we need from you

Five capabilities, in the order you'll use them.

1. Problem framing. You take a two-line business need from a client and turn it into a brief Claude can act on. The clarity of your writing determines the quality of the output.
2. Quality judgment. You read Claude's output and know whether it's done, draft, or wrong. You reject roughly 30% on first pass. That is the job, not a flaw in the system.
3. Iteration stamina. You can run the same loop fifteen times to get one thing right, and still be sharp on the sixteenth pass.
4. Cross-system thinking. You can hold a client's sales pipeline, calendar, inbox, billing system, and team chat in your head at once and see where the leverage point is.
5. Privacy and risk awareness. You instinctively know what data must never leave a tier, what action is irreversible, and when to ask before acting.

You have used Claude (or an equivalent frontier model) every day for at least six months. You have shipped automations that real people rely on. Your written and spoken English is strong enough to run a customer call without a translator. You have absorbed git, JSON, YAML, HTTP, and OAuth as a power user: not from a CS degree, and not by avoiding them.

You do not need to have written software for a living.

What we offer

• Direct CEO reporting and unfiltered context. You sit one chair from the decisions.
• A working blueprint, not a slide deck. You inherit a system already running in production. You do not have to invent it.
• A career trajectory that does not exist anywhere else. Successful candidates exit this role employable globally as senior AI-ops operators. The market for that capability is forming now; you will be at the front of it.

Package

Equipment stipend

USD $500 once-off for a working setup.

Claude credits

Fully covered by Seby.

Annual leave

15 days paid + Vietnamese public holidays.

Sick leave

7 days paid.

Working hours

Core overlap with Sydney business hours (4 hours / day minimum). The rest of your day is yours.

Location

Remote-first. Work from anywhere with a top-quality internet connection.