AI Video Pipeline Engineer (MiniMax Hailuo Specialist)

We are scaling our AI-driven video production pipeline using MiniMax H3 and are seeking a specialized engineer to design, optimize, and programmatically orchestrate our generation pipeline. The candidate will be responsible for balancing raw output resolution, batch cost efficiency, long-form shot continuity, and post-production audio integration. 1. Target Cost & Resolution Optimization 768p Resolution: Programmatically achieve and maintain 15-second outputs at less than $0.08 per clip through API orchestration, efficient batching, and cost-optimized compute infrastructure. Higher Resolutions (1K / 2K / 4K): Design and benchmark a hybrid pipeline (e.g., native 768p generation paired with real-time AI upscaler models like Real-ESRGAN or Topaz) to evaluate trade-offs between native render pricing vs. post-processed upscaling costs for 1K, 2K, and 4K deliverables. 2. Long-Form Video Orchestration (30s – 60s) Architect an automated workflow to seamlessly stitch 5–15 second native generations into coherent 30-to-60 second final videos. Implement frame-to-frame continuity strategies, utilizing first/last frame reference loops, latent state extensions, or omni-reference prompting to prevent character drift and visual degradation across multi-shot cuts. 3. Audio & Background Music (BGM) Pipeline Build a downstream audio-sync processing pipeline (e.g., via FFmpeg, Python/MoviePy, or API renderers) to ingest raw video outputs and merge them with dynamic BGM, dialogue, and sound effects. Leverage MiniMax H3’s native audio alignment or implement automated beat-matching scripts to synchronize camera cuts and key scene transitions directly to the musical cadence. Note to Applicants: Please provide a authentic, engineer-to-engineer response based on actual hands-on experience. Generic AI-generated applications or generic boilerplates will be filtered out.