AI Animation Quality Evaluation

About the project

I’m building a rigorous pipeline for reviewing and refining the short animated clips and visual-effects shots our generative models produce. To make that happen, I need someone who blends solid video & animation craft with hands-on AI and machine-learning experience. What you’ll actually do • Watch each AI-generated sequence, flag narrative, timing, or visual issues, and then suggest concrete fixes. • Co-design scoring rubrics and an evaluation framework that capture objective quality (resolution, frame consistency, colour, physics) and subjective storytelling criteria. • Curate a growing library of clearly labelled source files, reference renders, and metadata so future model-training datasets are instantly searchable. • Document edge cases and failure modes you discover, translating them into bite-sized tickets the engineering team can act on. • Provide concise written feedback after every iteration so the whole remote team can track improvements. Deliverables & acceptance 1. A version-controlled rubric (Markdown or Google Doc) complete with example clips for each score tier. 2. A folder of at least 200 annotated clips, each carrying your quality score, issue tags, and any replacement media. 3. A weekly report summarising key failure patterns, quick-win fixes, and longer-term research opportunities. All files must follow the naming convention we’ll share on kick-off, and every link in the report must resolve without extra permissions. You’ll collaborate asynchronously in our custom web dashboard, plus standard tools like Slack and Trello. Strong communication in written English is a must because most feedback happens in text. If you’re comfortable toggling between After Effects, Blender, or similar creative suites and jumping into model-training conversations about diffusion or transformer tweaks, I’d love to work with you. Good English skill required

Skills required