AI Song Recommendation Web App
Budget: ₹100 – ₹400 INR
I’m building a web application that will suggest songs in real time and I’d like a seasoned developer to take it from concept to a polished, deploy-ready product. The core of the system will be a Python + Flask backend that ingests song metadata and the ratings users give each track, runs them through a scikit-learn or similar model, and returns personalised suggestions. Pandas will handle data wrangling; any NLP that helps refine metadata or tag extraction is welcome.
On the front end, the interface needs to stay light but engaging: a responsive HTML/CSS layout with a clear section where listeners can create their own playlists and, inside each track card, click-through stars (or another intuitive widget) to register their manual ratings. Those ratings, alongside the underlying song attributes, should feed straight back to the model so the recommendations improve with every interaction.
Key deliverables
• Flask API connected to a trained recommendation model
• Clean, documented Python code (Pandas, scikit-learn, or compatible libraries)
• Responsive UI that supports playlist creation and in-card rating submission
• End-to-end testing script or notebook demonstrating that new ratings alter the suggested list appropriately
• Quick deployment guide so I can spin the app up on a VPS or Heroku-style platform
Acceptance criteria
The app must return at least five new, non-repetitive song suggestions within two seconds of a user rating a track, and those suggestions must demonstrably change after additional ratings are submitted.
If this matches your expertise, let’s discuss timelines and model choices so we can start shaping the music experience together.
On the front end, the interface needs to stay light but engaging: a responsive HTML/CSS layout with a clear section where listeners can create their own playlists and, inside each track card, click-through stars (or another intuitive widget) to register their manual ratings. Those ratings, alongside the underlying song attributes, should feed straight back to the model so the recommendations improve with every interaction.
Key deliverables
• Flask API connected to a trained recommendation model
• Clean, documented Python code (Pandas, scikit-learn, or compatible libraries)
• Responsive UI that supports playlist creation and in-card rating submission
• End-to-end testing script or notebook demonstrating that new ratings alter the suggested list appropriately
• Quick deployment guide so I can spin the app up on a VPS or Heroku-style platform
Acceptance criteria
The app must return at least five new, non-repetitive song suggestions within two seconds of a user rating a track, and those suggestions must demonstrably change after additional ratings are submitted.
If this matches your expertise, let’s discuss timelines and model choices so we can start shaping the music experience together.