rank-design

Designs ranking pipelines with reranker selection, score fusion, and cross-encoder patterns. Useful for optimizing search or recommendation ranking with latency trade-offs.

Designs recommendation/ranking pipelines using the six-stage Source→Hydrator→Filter→Scorer→Selector→SideEffect pattern. For social feeds, RAG rerankers, ad ranking, and any top-K selection system.

Builds ranking evaluation workflows including NDCG/MRR metrics, human relevance labeling, and offline evaluation harnesses. Useful for search relevance and recommender system testing.