recsys-pipeline-architect

Name: recsys-pipeline-architect
Availability: InStock
Author: Canhada-Labs

Design recommendation systems using a six-stage pipeline architecture.

2|Updated Jun 30, 2026

Search Tags:#ranking #pipeline architecture #recommendation system #recsys #feed algorithm

Ready to run

npx skills add https://github.com/Canhada-Labs/ceo-orchestration --skill recsys-pipeline-architect-canhada-labs

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: recsys-pipeline-architect
Source: https://github.com/Canhada-Labs/ceo-orchestration/tree/main/.claude/skills/domains/architecture/skills/recsys-pipeline-architect
Command: npx skills add https://github.com/Canhada-Labs/ceo-orchestration --skill recsys-pipeline-architect-canhada-labs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill helps users design and build efficient recommendation systems by providing a structured pipeline architecture, guiding the creation of composable recommendation, ranking, and feed pipelines.

Core Features & Use Cases

Structured Pipeline Architecture: Offers a six-stage pattern for building recommendation systems: Source → Hydrator → Filter → Scorer → Selector → SideEffect.
Stage Order and Trade-offs: Explains the importance of stage order and trade-offs such as single score vs multi-action prediction, candidate isolation vs joint scoring, and online vs offline vs hybrid serving.
Interview-to-Scaffold Workflow: Provides an eight-step workflow for converting a use case into a working pipeline, including clarifying the use case, enumerating candidate sources, listing required hydrations, and more.
Cookbook and Rules: Includes a filter and scorer cookbook, hard rules for best practices, and anti-patterns to avoid.

Quick Start

Use the recsys-pipeline-architect skill to create a recommendation system pipeline for a social feed.

Frequently Asked Questions about recsys-pipeline-architect

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I design a recommendation system pipeline architecture for a social feed?▼

A recommendation system pipeline architecture uses a six-stage pattern: Source, Hydrator, Filter, Scorer, Selector, and SideEffect. This structured approach helps build composable feed pipelines by defining stage order and trade-offs like online versus offline serving.

What are the stages of a recommendation pipeline and how do I order them?▼

The stages of a recommendation pipeline are Source, Hydrator, Filter, Scorer, Selector, and SideEffect. Correct stage ordering is critical for managing trade-offs such as single score versus multi-action prediction and candidate isolation versus joint scoring.

How do I convert a recommendation use case into a working feed algorithm pipeline?▼

To convert a recommendation use case into a working pipeline, follow an eight-step workflow. This process includes clarifying the use case, enumerating candidate sources, and listing required hydrations to scaffold a functional feed algorithm pipeline.

What are common anti-patterns and best practices when building ranking pipelines?▼

When building ranking pipelines, adhere to hard rules for best practices and avoid specific anti-patterns. The pipeline architecture provides a filter and scorer cookbook to guide single score versus multi-action prediction decisions and prevent common structural mistakes.

Can I use this pipeline architecture for both offline and online recommendation serving?▼

Yes, the pipeline architecture supports online, offline, and hybrid recommendation serving. It explicitly addresses trade-offs between these serving modes to help you structure composable ranking and feed pipelines efficiently.

Does building a recsys pipeline require prior knowledge of pipeline patterns?▼

Yes, building a recsys pipeline requires knowledge of the six-stage pattern and the ability to implement a pipeline according to provided guidelines. This prerequisite ensures you can correctly apply stage order and structural trade-offs.