Vespa Recommendations guides for WordPress & WooCommerce

Vespa.ai is the perfect solution for personalized search and recommendations with XGBoost or LightGBM models, but it requires external code to create the models. Metarank Labs is the perfect no-code solution for training LTR models, thanks to its predefined feature store recipes, but it lacks the flexibility of pure vector search to provide perfect accurate results. So we propose to combine the two as the first no-code Learning to Rank solution that supports the full cycle: Ingest items in Metarank and Vespa Ingest events in Metarank Transform items and events with feature store recipes as yaml parameters Train models on the feature store Export the model to Vespa.ai as an internal ONNX embedder Rank/rerank the search/recommendations with ranking expressions including the trained model