recommendation-ml

What problem does it solve?

This Skill helps data teams and engineers rapidly develop and deploy ML-driven recommendation systems for e-commerce, reducing time-to-value from model to production.

Core Features & Use Cases

• End-to-end recommendation pipelines: Collaborative filtering, content-based, and hybrid models with scalable feature stores and registry.
• Production-grade tooling: Feast feature stores, MLflow model registry, model evaluation, and governance.
• Experimentation and diversity: Thompson Sampling, MMR-based diversification, A/B testing, and cold-start strategies for new users/products.
• Use Case: Build a complete recommender for an online store, from data ingestion to serving real-time recommendations.

Quick Start

Start with a baseline model by following included examples to train a matrix factorization model on your product dataset and register it with MLflow.