Leveraging social interactions and AI to increase engagement, gather valuable user insights, and optimize ad delivery on the Joyn platform.
Joyn Social is a proof-of-concept platform built during a one-week hackathon hosted by 7.1 Tech Hub, a ProSiebenSat.1 Media SE company. Our goal was to develop innovative ways to:
- Increase user engagement,
- Collect more user data (especially for cookie-less users),
- Improve content recommendation using AI,
- Enhance ad targeting capabilities.
We designed a social layer on top of the Joyn streaming experience—bringing people closer together through shared content while improving business value through smarter recommendations.
- Add friends via Joyn Social
- Get better content recommendations based on your friends' watch history
- Create, share, and explore public watchlists
- Suggest shows to friends based on mutual activity
- Tinder-style swipe interface for shows/movies
- Swipe right to like (adds to watchlist), swipe left to dislike
- Builds personal taste profile
- Feeds directly into AI recommendation models
We used a hybrid AI recommendation approach:
- Collaborative Filtering: Finds similar users and recommends what they liked
- Content-Based Filtering: Uses metadata from media (e.g., genre, actors, tags)
| Problem | Our Solution |
|---|---|
| 🍪 40% cookie-less users | Use friend activity and watchlists to infer preferences |
| 🎯 Ad Targeting | Leverage friend network proximity to optimize ad delivery |
| 📈 Increase engagement | Introduce social interactions and gamify content discovery |
| 🚀 Reach new users | Shared watchlists and recommendations generate viral potential |
- Frontend: React.js (prototype UI for testing features)
- Backend: Node.js (API server)
- Database: SQL database hosted on AWS RDS
- AI Models: Python (Scikit-Learn / Sentence-transformers)
- Cloud Infra: AWS (RDS, EC2 for model training/deployment)
We measured our AI performance using:
- ✅ CTR (Click Through Rate)
- ✅ Useful Recommendations per Recommendation
- ✅ NDCG (Normalized Discounted Cumulative Gain)
This evaluation allowed us to fine-tune our model and deliver highly personalized content suggestions with measurable impact.
- Integration with real Joyn accounts
- Fine-tune ad recommendations based on collective engagement
- Scaling the model
This project was created as part of the 7.1 Tech Hub Hackathon.
Made with ❤️ by a passionate team of developers!
Have questions or want to collaborate? Reach out via linkedin or email

