{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/7010f453ab2c4b479ec040bdbc60a63a\" 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/7010f453ab2c4b479ec040bdbc60a63a-cf4919a448293e2a.gif","duration":535.977,"title":"Spotify Music Recommender with LLM Parsing 🎵","description":"For my AI 1.10 final project, I expanded a music recommender from about 17 songs to around 210 songs using my Spotify playlist. Users type a natural sentence, then an LLM parser returns structured preferences like genre, mood, energy, acousticness, and related factors, and a recommend function scores and retrieves matching tracks. I improved search by allowing multiple genres per song, and I added language as part of the genre metadata. I tested cases like 12 reggaeton and bachata songs and an even split of 5 Spanish and 5 Chinese songs, and found limitations like uneven splits and no conversational history. I did not request any action from viewers."}