recommendation-systems-engineering

v0.1.2

Recommendation-systems-engineering team — the RECSYS layer that turns interaction data into ranked recommendations. agents (recsys-architect, recsys-implementation-engineer) covering approach selection (popularity baseline, collaborative filtering, content-based, hybrid, two-tower/embedding retrieval, sequential), the candidate-generation → ranking → re-ranking pipeline, cold-start (new user/new item), offline evaluation (recall@k, nDCG, MAP, coverage/diversity) AND online A/B, feature stores with train/serve parity, low-latency serving, and feedback-loop/position bias. skills, Mermaid-backed knowledge docs, templates. Seams: training infra → ml-engineering; keyword/semantic search → search-relevance-engineering; A/B + stats → experimentation-growth-engineering + applied-statistics. House line: baseline before a neural net; offline wins must survive an online A/B. Requires ravenclaude-core@>=0.7.0.

Claude Code3 Skills

By Matt CorbettLicense: MIT7 GitHub starsUpdated last week

Directory evidence

Runtimes: Claude Code
Parsed components: 3 skill or MCP entries
Source updated: Sep 15, 2026
Manifest status: Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install recommendation-systems-engineering for Claude Code

Installs for the current user

claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install recommendation-systems-engineering@agent-plugin-marketplace

Paste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually

$ git clone https://github.com/mcorbett51090/RavenClaude

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/recommendation-systems-engineering/.

Plugin files

plugins/recommendation-systems-engineering/

├── .claude-plugin/plugin.json
├── skills/choose-recsys-approach/SKILL.md
├── skills/evaluate-recommenders/SKILL.md
└── skills/handle-cold-start-and-serving/SKILL.md

Included Skills3

choose-recsys-approachskills/choose-recsys-approach/SKILL.md

Choose the recommendation approach — popularity/heuristic baseline, collaborative filtering, content-based, hybrid, two-tower/embedding retrieval, or sequential — from data volume, sparsity, cold-start severity, latency, and interpretability. Use when starting a recsys or deciding whether to add model complexity.

evaluate-recommendersskills/evaluate-recommenders/SKILL.md

Evaluate recommenders correctly — a temporal train/test split, per-stage offline metrics (recall@k, precision@k, nDCG, MAP, coverage, diversity, novelty) against a popularity baseline, AND the online A/B that is the real verdict. Use to build an eval harness or diagnose an offline-vs-online gap.

handle-cold-start-and-servingskills/handle-cold-start-and-serving/SKILL.md

Handle cold-start (new users and new items) and serve recommendations within a latency budget — content/popularity fallbacks, onboarding & exploration, ANN retrieval, precompute/caching, online feature fetch, and popularity-on-timeout fallback. Use when new entities recommend poorly or serving is too slow.

Plugin manifests1

plugins/recommendation-systems-engineering/.claude-plugin/plugin.json

{
  "name": "recommendation-systems-engineering",
  "version": "0.1.2",
  "description": "Recommendation-systems-engineering team — the RECSYS layer that turns interaction data into ranked recommendations. agents (recsys-architect, recsys-implementation-engineer) covering approach selection (popularity baseline, collaborative filtering, content-based, hybrid, two-tower/embedding retrieval, sequential), the candidate-generation → ranking → re-ranking pipeline, cold-start (new user/new item), offline evaluation (recall@k, nDCG, MAP, coverage/diversity) AND online A/B, feature stores with train/serve parity, low-latency serving, and feedback-loop/position bias. skills, Mermaid-backed knowledge docs, templates. Seams: training infra → ml-engineering; keyword/semantic search → search-relevance-engineering; A/B + stats → experimentation-growth-engineering + applied-statistics. House line: baseline before a neural net; offline wins must survive an online A/B. Requires ravenclaude-core@>=0.7.0.",
  "author": {
    "name": "Matt Corbett"
  },
  "homepage": "https://github.com/mcorbett51090/RavenClaude",
  "license": "MIT",
  "keywords": [
    "recommendation-systems",
    "recsys",
    "collaborative-filtering",
    "content-based-filtering",
    "two-tower",
    "candidate-generation",
    "ranking",
    "re-ranking",
    "cold-start",
    "ndcg",
    "recall-at-k",
    "embeddings",
    "feature-store",
    "online-evaluation",
    "position-bias"
  ],
  "requires": {
    "plugins": [
      "ravenclaude-core@>=0.7.0"
    ]
  }
}

For maintainers

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[recommendation-systems-engineering on Agent Plugins Marketplace](https://pluginsmp.com/plugins/recommendation-systems-engineering)