Completed a ~90-second cinematic micro-drama assignment for an AI artist role, working from rough scene outlines, simple dialogue lines, an audio track to sync to, and provided character and location reference images. Engineered detailed, shot-by-shot prompts for Google Flow to generate 13 continuous live-action clips in 9:16 vertical format, translating the simple source material into precise continuity instructions covering consistent character appearance, wardrobe, and environment, physically accurate prop handling (a cigar-and-lighter sequence tracked frame-by-frame from pocket to lit cigar to ash), and natural camera movement. The generated clips were then stitched together and synced to the provided audio to produce the final scene.