x-algo-pipeline

What problem does it solve?

Clarifies how the X recommendation pipeline processes and ranks posts from hydration to final feed, enabling quick debugging, optimization, and capacity planning for large-scale recommendations.

Core Features & Use Cases

• End-to-end eight-stage pipeline overview (Query Hydration, Sources, Candidate Hydration, Pre-Score Filtering, Scoring, Selection, Post-Score Filtering, Side Effects)
• Insight into how in-network Thunder and out-of-network Phoenix sources contribute to candidates
• Use cases: debugging ranking behavior, performance tuning, feature experimentation, and scalability planning

Quick Start

Trace the eight stages of the pipeline from a sample user request to the final feed output to observe where data is hydrated, scored, filtered, and selected.