Détails de l'annonce
Référence
247153
Date de publication
18 septembre 2026
Type de contrat
CDI SIVPLieu de travail
El Menzah, Tunis, Tunisie
Expérience requise
Entre 1 et 2 ans
Niveau d'études
Bac + 3
Salaire proposé
Disponibilité
Plein temps
Langues
Arabe Anglais
Entreprise
BUZZBIBLE
Secteur: culture / audiovisuel
Taille: Moins de 20 employés
Description de l'annonce
About the role
You write stories that get made by an AI agent rather than a crew. The whole pipeline runs through one person, and that person is you: the film starts as your idea, and every frame of it is something you described out loud.
The work has three movements. First you write the story — beginning, middle, end, characters worth watching. Then you break it into parts the generation tools can actually produce, each one a self-contained beat. Then you talk each part through with the AI agent, aloud, at the level of detail a producer gives an actor on set: where they stand, what the hands do, how the expression shifts on the third word, what the camera is doing while it happens.
That third movement is the job. Plenty of people can write a story. Far fewer can describe two seconds of it precisely enough that a model renders what was in their head — and then keep doing it aloud, in fluent English, for most of a working day. That specific skill is what we're hiring for.
Responsibilities
1. Write the story
- Turn a brief into a complete narrative — premise, characters, arc, ending
- Build it to the length and format the project needs
- Get it agreed before a single frame is generated
2. Break it into parts
- Divide the story into segments the generation tools can hold in one shot
- Keep continuity across segments: same faces, same wardrobe, same light, same world
- Decide where the cuts fall and what each part has to carry
3. Direct each part aloud
- Describe every segment to the AI agent vocally, in continuous, precise English
- Direct the performance the way a producer directs an actor: posture, gesture, eyeline, the timing of a reaction, the exact change in an expression
- Specify the camera in the same breath — framing, lens character, movement, where the shot opens and where it lands
- Specify light, texture, wardrobe, setting and sound so nothing is left for the model to guess
4. Review and re-direct
- Watch what comes back and name exactly what's wrong with it
- Re-describe until the output matches intent, and recognise when a part needs rewriting rather than re-describing
- Record the phrasings that work, so the studio's direction language sharpens over time
5. Deliver
- Hand the assembled parts to edit and sound with the notes they need
- Present and defend the work in English, to the team and to clients
Requirements
Spoken English, C1 or above. You'll be describing scenes aloud for much of the working day, in long unbroken stretches. Your speech has to stay clear, fluent and well-ordered in real time — not rehearsed, not searching for words. We test this directly: you'll direct a scene out loud in the interview.
Written English at the same level. The story, the breakdown and everything a client sees are written in English.
Storytelling. You can build a full narrative from a one-line brief, and you can say why an ending lands.
A producer's vocabulary for physical detail. You can describe a body, a gesture, a reaction and a camera move in words precise enough that two people would picture the same thing. This is the rarest part of the job and the one we weigh most heavily.
A cinematic eye. Shot size, lens character, lighting, blocking, pacing — you direct for them and you have the references to point at.
Hands-on experience with AI video generation. You've used the tools on real work. You can show what you made, what failed first, and how you talked your way from one to the other.
Patience with iteration. Generated video rarely lands on the first pass. You can run the same six seconds twenty times without losing the thread or the standard.
Experience. [1–4] years writing or directing video, branded content, short film or generated content. A portfolio that proves the skill counts more than the years. No particular degree is required.
Nice to have
None of these are required. Any of them will move your application up.
- A performance background — acting, directing actors, dubbing or radio. People who have directed a human find it far easier to direct a model.
- Editing experience, so you break the story into parts that actually cut together.
- Storyboarding, by hand or in software.
- Sound instincts — you hear the scene as well as see it.
- A personal body of generated work — a channel, a series, shorts you made because you wanted to.