Description
Summary I’m looking for an experienced AI/computer vision engineer to build an internal tool that helps us generate higher-performing YouTube podcast thumbnails. We produce long-form interview podcasts. Our thumbnail design is highly standardized (same overall layout, fonts, branding, etc.). The primary variable is the facial expressions of the host and guest. The goal is to automatically analyze a podcast recording, extract the strongest facial expressions from both participants, and generate multiple thumbnail candidates optimized for click-through rate (CTR). This is an internal production tool, not a public SaaS product. Will send more details to top freelancers. In your proposal tell me about relevant related projects you’ve completed.