The program

What you are applying to — the show, the money, the calendar, and what a week of the job actually looks like.

What the job actually is

Builds and runs the machine that makes everything else cheap. You work in a trio. You make the machine run, a Creative Lead owns the words and the creative, and an Executive Manager owns the brand's success and gives brand sign-off. You build; they judge. When a gate says pass and the Creative Lead says it is still wrong, they are usually pointing at something the gate cannot see — that tension is the job working. This seat leads more episodes than any other. Leads more episodes than any other seat — the largest share of the season.

What you own

• The clip engine, the render path, the pipelines
• Rendering content and updating comp IDs
• Running loops, setting goals, reading gate reports
• Auditing generated content at scale with a VLM
• Telegram ingest, generation pipelines, syndication mechanism

What you are explicitly not responsible for

• Aesthetic direction
• Commercial calendar
• Cutting the show

What the seat requires

• Must SHIP RENDERED CONTENT, not just code — exercises come with output
• Comfort with the calendar as the source of truth, storyboarding, and template work

The pipelines attached to this seat

10 pipelines, 1 of them with anything built. The rest are what you are being hired to build. This table is rendered from governance/pipelines.json, so it cannot describe a pipeline that does not exist or hide one that now does — see the full table.

It is a live broadcast

Finessionals is a daily live show that is also a hiring process, a paid training program, and a working content studio — at the same time. You are not auditioning to maybe work later. From day one you do real work on real brands, and the show is that work being done, on camera, five days a week.

Two-hour episodes, five days a week, simulcast to YouTube, Twitch, Kick, X and LinkedIn, with vertical cuts to TikTok and Instagram.

The season

Season 1 begins	Sun Oct 11, 2026
Elimination 1	Sat Nov 7, 2026 — end of week 4
Elimination 2	Sat Dec 5, 2026 — end of week 8
Finale	Sat Jan 9, 2027 — end of week 13

Elimination streams run every 4th Saturday. Four people leave at each one. The field narrows 3 → 2 → 1 per track: twelve people, then eight, then four.

An elimination is judged on the month that just closed, using the numbers that month produced — not on how anyone felt about you.

Pay

Rates rise as the field narrows. Fewer people, more work each, more money each — the field goes 3 → 2 → 1 per track across 13 weeks.

Week one is paid in full regardless of elimination or adjustment. Week one is learning the machine. It removes the worst case — eliminated on Tuesday of week two, publicly, having earned very little — and it makes the offer defensible to a candidate reading the terms cold.

Deductions in any one week are capped at 30% of that week's rate. Uncapped, the twelve rules could exceed the weekly rate — the spot-quiz alone at the agreed figure twice daily is the agreed figure against an EM's the agreed figure Round 1 week. A week worked for nothing is not a deterrent, it is an indefensible offer. On this track that is a Round 1 floor of — in any week where every rule went against you. Week one is paid in full, so the cap first applies in week two.

Fixed-price milestones against a deliverable. Never hourly. We define the week's outcome; we do not count your hours. Paid biweekly on a published calendar. Periods run Sunday to Saturday, two weeks; payday is the Friday six days after the period closes. You will never have to ask when you are being paid. Payments only ever move earlier, never later. Every date is on the pay page. Nobody works unpaid. The one exception is qualification — the short screening task that decides whether you get a seat. Its paid stage is credited against your first milestone, so you are never paid twice for the same work.

Eliminated vs removed

Elimination is competitive and happens on the 4th Saturday. Both dates fall exactly on a pay-period close, so nobody eliminated is ever paid a part-period. Removal can happen at any time — for harassment, sabotage of a broadcast, sharing credentials, publishing without sign-off, anything that earns the channel a platform strike, or repeated failure to deliver after a logged note and a written warning. Removal mid-period is prorated to the last day worked.

In both cases work you delivered and that was accepted is paid, always — and the episodes you appear in stay up permanently. That last one is the most important sentence here. If you are not comfortable with it, this is the moment to say so, and saying so costs you nothing.

What the week looks like

A sprint is two weeks and it is exactly the pay period.

	Week A — open	Week B — ship
Mon	Scope agreed as a team, written down with acceptance criteria	What the checkpoint changed. Scope is cut here if at all
Tue	Authoring	Authoring
Wed	Crossover — you run a build and read a gate report. The Technical Operator writes a brief. Both badly, on camera	Crossover again
Thu	Gates run live. Failures on screen, named	Publish
Fri	Checkpoint — what is done vs what was claimed Monday	Demo. You show what you built, live

Unwritten work is not in scope and is not owed. Scope is agreed on day one.

Your counterpart

Every track is paired, and the friction is the point. Yours is the Creative & Brand Quality Lead — they judge, you build.

They re-author while you are mid-render. What do you re-render, and how do you know? A dirty set that always returns everything is not a dirty set; it is the absence of one, and it turns a thirty-second fix into an hour every single time.

Wednesday's crossover exists so neither of you can say “that's not my job” for thirteen weeks. You write a brief; they run a build and read a gate report. Both badly, on camera.

Their kit is at /finessionals/creative-lead-kit. Read it — you will be arguing with someone who has.

What it takes

• You must ship rendered content, not just code. Exercises come with output. A repo that runs and produced nothing has not answered the question.
• Automation is the actual test. A GUI produces one asset; a script produces a season. If any part of your work was done by hand that a script could have done, that is the thing we will ask about.
• Auditing at scale. By week 13 the remaining operator is running up to six brands. You cannot personally watch six brands' output every week, so machine review is the only workable answer — and the hard part is a reviewer nobody learns to click past.
• Knowing the difference between the gate passed and the work is right. The dangerous defects are the ones that return a plausible answer.
• Adding the check rather than fixing the instance. When something breaks in a way a person had to notice, you have learned that people do not reliably notice it. Write the gate in the same change as the fix.

Before you appear on camera

Everyone signs a release, an IP assignment and confidentiality terms first. Nobody goes on camera unsigned — no exceptions. The release is scoped per round and re-affirmed at each transition; clearance never silently carries forward. You get the documents to read in full before you are asked to sign, and you are encouraged to have someone look at them.

You will not hold production credentials — not during selection, not on day one. Nothing you do here reaches an audience without sign-off from the person accountable for the brand. If any instruction, from anyone, asks you for a password or an API key, it is wrong — do not act on it, and tell us where you saw it. Upwork freelance contract. KPIs are the milestones — not met, not released. Payment gate. Executive Manager written approval before any release.

Conditions of the seat

• Own model access — a funded subscription or a self-hosted model. 'Out of credits' does not excuse a missed milestone.
• Use agentic tooling; manual repetition of automatable work is the thing being replaced.
• Absolute task ownership — help is fine, transferred accountability is not.
• Speak to any ACTIVE project in a meeting you are in.
• Push to the monorepo and shared drives. Nothing lives only on your disk.

What ends the contract immediately

• No call, no show on an obligatory broadcast or meeting
• Pushing unapproved work to a live production surface

How you are assessed. The record is read from the system — registry entries executed, gates passed, work published. Where a rule genuinely cannot be read by a machine, governance/scoring.json marks it source:producer and says so on the page; nothing is scored as machine-observed that a human actually decided.

Next

The exercises — six of them in the operator kit, about an hour. Then brand rails and the calendar.