AI Creator
AI stills and AI video in one pipeline, with one recognizable look across every format.
Position summary
You own AI content for short-form, YouTube and social, from brief to finished asset. Stills and video live in one pipeline: key visuals, AI avatars, generation series, image-to-video, lip sync, editing. Underneath runs multi-model production across Midjourney, ComfyUI, Higgsfield, Runway, Kling, Seedance and whatever ships next.
Content usually breaks at the seam between formats: the key visual in the feed does not match the hero in the video, lip sync drifts, lighting jumps, and by the second release the series has fallen apart. Your job is one recognizable look across stills and video. The Brand Visual Designer sets the look and direction; tasks and acceptance come from the founder and the Head of Ops.
Key responsibilities
- Static assets in consistent series: key visuals, AI avatars, covers, carousels, infographics, icons, mockups.
- Prompt systems for stills: text-to-image, image-to-image, style replication from references and guidelines.
- AI video from scratch: keyframes with a consistent hero, image-to-video, lip sync, multi-shot scenes with continuity.
- Long videos assembled from short AI clips: camera, motion strength, transitions between scenes.
- Editing and sound mix in DaVinci Resolve or Premiere, synced with ElevenLabs and team voiceover.
- AI-native documentation: specs, shot lists, asset versions. Generation starts once the scope is defined.
Required qualifications
Read this block before applying. If most of these are a no, we won't be able to take you on this role, and we'll both lose the time.
- 3+ years in visual production, motion or AI-native content production.
- Both stills and video: AI image series plus AI character animation for YouTube or short-form, not one or the other.
- You hold hero identity across formats: keyframes, lip sync and continuity of wardrobe, lighting and background.
- Prompt engineering for consistent series: image-to-image, text-to-image, style replication from references.
- You work confidently in image-to-video and text-to-video and assemble long video from short clips.
- Multi-model workspaces: Higgsfield or equivalents such as Luma, Pollo AI, or aggregators with Kling, Veo, Sora and Seedance in one interface.
Nice to have
- ComfyUI with LoRA training on Flux or Stable Diffusion for custom characters and brand style.
- Google Veo, OpenAI Sora or Wan, via Higgsfield or direct access.
- After Effects: keyframes, masks, animated typography.
- fal.ai or API access to models for custom pipelines.
Stack
- Stills: Midjourney, Krea, ComfyUI, LoRA on Flux and Stable Diffusion
- Video: Higgsfield (Kling, Veo, Sora, Seedance, Wan), Runway Gen-3 and Gen-4, Kling 3.0, Seedance 2.0
- Audio and avatars: ElevenLabs, HeyGen
- Editing and motion: DaVinci Resolve or Adobe Premiere, After Effects
Success metrics
This is what the work in this role is measured against, and what we'll walk through at the final interview.
- Stills: 2–3 assets per account per week — key visuals, infographics, carousels, covers.
- AI avatar experiments: about five Reels per week across two or three new products.
- The founder's Instagram: three to four AI videos per week.
- YouTube: one longer piece per week plus three or four Shorts.
- Look consistency: few revisions on continuity and brand fit after the first agreed scope.
Employment terms
- Format
- Full-time, remote. Flexible hours, sync with the team on CET.
- Compensation
- From $4,000/mo fixed (USDT)
- Team
- Direct work with the founder and domain leads
- Network
- Direct work with Silicon Valley funds and the studio founders
What we offer
- A fixed base set by role and stream, plus bonuses for shipped releases and metrics hit.
- Revenue share on the product or stream and options in the pool. You earn like a co-owner.
- A fast track earned by results: more products, more budget, more equity.
- Remote-first, flexible hours, core overlap 15:00–19:00 CET. Work from anywhere.
- Direct access to the founders and an AI-native stack with a budget. No five layers of approvals.
How the process works
We hire slowly and deliberately. Every step is about keeping it honest and aligned for both sides.
Screening
We read every application ourselves, no ATS black box. Then a short call with the hiring team to align on experience, expectations and motivation.
Deep interview
A long conversation with the stream lead about your track record, real cases and how you decide under uncertainty.
Culture fit
One-on-ones with the core team and a founder. We look at values, ethics and how you behave when things get hard.
References
We talk to people you've actually worked with. Hygiene for us, not a formality, and we always do it.
Pilot contract
Two to four weeks on a real product with a clear Definition of Done and concrete artifacts. We don't give take-home tasks.
A 30/60/90 plan and the long game
We match results both ways, set a plan for the first three months and go long. Or part ways honestly.