Recommendation Engine Setup
by Community · free · Last verified 2026-03-17
This script provides a complete setup for a modern, two-stage recommendation engine. It uses a two-tower neural network for efficient candidate retrieval and a powerful Large Language Model (LLM) for nuanced re-ranking. The system integrates with a Feast feature store to leverage real-time user context, ensuring timely and relevant suggestions.
https://github.com/NVIDIA-Merlin/Merlin ↗C+
C+—Average
Adoption: BQuality: AFreshness: ACitations: C+Engagement: F
Specifications
- License
- MIT
- Pricing
- free
- Capabilities
- Two-tower model for scalable candidate generation, LLM-based cross-encoder for sophisticated item re-ranking, Real-time feature serving via Feast feature store integration, Content-based fallback mechanism for cold-start users and items, Built-in scaffolding for A/B testing and algorithm experimentation, Personalized recommendations based on user interaction history, Hybrid approach combining collaborative and content-based filtering, Scalable architecture for handling large item catalogs and user bases
- Integrations
- Feast, TensorFlow, PyTorch, Hugging Face Transformers, Apache Spark, Kubernetes, MLflow
- Use Cases
- [object Object], [object Object], [object Object], [object Object], [object Object]
- API Available
- Yes
- Language
- python
- Dependencies
- torch, feast, faiss-cpu, fastapi, redis, numpy
- Environment
- Python 3.10+, CUDA optional
- Est. Runtime
- Training: 1-4 hours; inference: <20ms
- Tags
- recommendation-engine, collaborative-filtering, llm-reranking, two-tower-model, personalization, feature-store, feast, deep-learning, mlops, ab-testing, retrieval-ranking
- Added
- 2026-03-17
- Completeness
- 0.8%
Index Score
58.7Adoption
68
Quality
85
Freshness
84
Citations
58
Engagement
0