{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/837b4819f0db4fb288351536277cd294\" frameborder=\"0\" width=\"1728\" height=\"1296\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1296,"width":1728,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1296,"thumbnail_width":1728,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/837b4819f0db4fb288351536277cd294-000eb48adcb641cf.gif","duration":580.289,"title":"Wayfinder RAG Music Recommender Project","description":"Hi, I am Rishi, and this is my CodePath AI final project submission, Wayfinder, a music recommender powered by RAG and Gemini. I built a system using Kaggle track data with attributes like energy, tempo, danceability, valence, and artist and track names, then created embeddings plus a TF-IDF matrix, and used them with ChromaDB for hybrid semantic and keyword retrieval. I demoed prompts like high energy workout tracks and chill indie music and showed match scores and an AI summary that hand picks top tracks without hallucinating. I did not request any action from viewers, but I included how to run it locally from my Git repo and set a Gemini API key in the environment file."}