{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/c2f54c6782a545cb978504ccfa60246a\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/c2f54c6782a545cb978504ccfa60246a-7350b732e2e9d86d.gif","duration":238.059,"title":"Corkaminder Music Recommender, Rule and RAG 🤖","description":"Hi, I am Tessie, and this is my final project, the Corkaminder system. I show two parts, a rule based swear that scores songs to a user taste profile using weighted rules, and a RAG pipeline that uses TF IDF retrieval and an AI model to generate recommendations. I demo multiple profiles like Late Night Jazz, show an edge case where classical music is missing from the catalog, and a mismatch case where no classical songs were retrieved and the AI still suggested pop and ambient tracks honestly. I also test guard rails for missing API configs and logging, and one aggressive metal profile returned an acoustic track but noted low acousticness 0.07 and higher aggressiveness and energy. No action was requested from viewers."}