AI Video Generator + Editor
๐ Philippines
๐ฐ $3-4/video
๐
Posted 2026-08-20 ยท via OnlineJobs
๐ท remote
Apply on original site โASSIGNMENT FOR THE JOB. PLEASE REVIEW THE JOB DESCRIPTION AND IF YOU ARE A GOOD FIT PLEASE COMPLETE THIS SHORT RECREATION ASSIGNMENT.
Use the reference photo: ----------
And recreate this video using our character from the reference above: ---------- ==
Ensure the frames all look very similar but make sure you adjust the background, dog and man to fit our character. Also slightly change the script to fit huskies rather than pit bulls.
THIS ASSIGNMENT IS A MAJOR FACTOR IN THE HIRING PROCESS
JOB DESCRIPTION:
About the role
- Our team is looking to hire a new video editor! We produce short-form AI-generated video content, generated imagery, generated footage, and cloned voice, cut into finished pieces for social. This role owns that pipeline end to end: you generate the stills, animate them, clone and direct the voice, and cut the final edit.
- This is a maker role, not a strategy role. You won't be staring at a blank page.
- You'll be given a script and reference videos for every piece, so the creative direction is already set. Your job is to execute it well and fast. Volume and consistency matter more than any single hero video. If you like sitting in the generation loop, rejecting 90% of what comes out, and finding the 10% that actually looks real, this is your job.
What you'll be doing
Image generation
- Build the keyframes every video starts from, we generate stills first, fix them until they're genuinely convincing, then animate. No text-to-video shortcuts.
- Maintain character and identity consistency across scenes, angles, and lighting setups using reference images and trained identity models.
- Prompt with intent: match a specific key light, color palette, lens feel, and wardrobe rather than accepting whatever the model defaults to.
Video generation
- Animate approved keyframes in Google Flow, image-to-video, reference-driven generation, scene extension, camera motion.
- Work in short takes. Identity holds for roughly 5 seconds and drifts after, so the job is cutting more and generating less.
- Run and log iterations efficiently: know when a prompt is one nudge from working and when to throw it out and re-shoot the keyframe.
Voice
- Build and maintain voice clones in ElevenLabs, and manage the source recordings that feed them. (VOICE CLONES FOR ALL CHARACTERS WILL BE PROVIDED)
- Direct delivery pacing, emphasis, emotion โ and regenerate until a line reads naturally instead of like narration.
- Clean up audio, match lip and body timing to the visuals, and keep voice consistent across an entire series.
Editing & finishing
- Cut everything together in CapCut: pacing, hooks, transitions, captions, sound design, music beds.
- Color and grade-match clips from different generations so a sequence reads as one continuous piece.
- Export and upload to google drive for review
Scripting & workflow
- All scripts and reference videos will be provided daily so you don't have to stress about what to create
- If accepted into this job you will be provided access to a tool that improves your workflow and speed drastically (this means you make more $$)
Quality control
- Catch what breaks the illusion before we publish: identity drift, over-smoothed skin, mismatched light direction, hand and face artifacts, warped background detail.
- Flag anything that creates a rights, likeness, or platform-policy problem.
What you'll be working from
- You get direction daily, this is not a role where you're left guessing what to make.
- Scripts are written and provided for you. You may refine wording for pacing or delivery, but you're not starting from scratch.
- Reference videos are provided for every piece, showing the look, pacing, tone, and structure we're going for. Your job is to hit that reference, then improve on it where you can.
- Feedback and revisions come regularly. Expect notes, and expect to turn them around quickly.
IMPORTANT: Coming in, you should be able to watch a reference video and break it do