skills.rest
hung-phan
hung-phanCOMMUNITY
Explore all skills by this author →
GitHub

recommendation-system

Design two-stage recommender systems with candidate generation and ranking components.

Updated Jun 5, 2026
Search Tags:#system design#personalization#machine learning#data engineering#recommender systems
One-click install
npx skills add https://github.com/hung-phan/system-skills --skill recommendation-system-hung-phan
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: recommendation-system
Source: https://github.com/hung-phan/system-skills/tree/main/skills/system-review/references/interview-templates/recommendation-system
Command: npx skills add https://github.com/hung-phan/system-skills --skill recommendation-system-hung-phan

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires sklearn, numpy, pandas, tensorflow, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides a structured framework for designing and implementing large-scale recommender systems, addressing the complexities of personalizing user experiences across platforms like YouTube, Netflix, and TikTok.

Core Features & Use Cases

  • Interview Template: Offers a structured approach to designing large-scale recommender systems suitable for interview settings or practical design challenges.
  • Two-Stage Architecture: Splits the system into candidate generation and ranking stages, optimizing for recall and precision respectively.
  • Feature Stores and Training: Utilizes feature stores for managing data efficiently and applies techniques for offline training and online serving to mitigate skew.
  • Cold Start Solutions: Provides strategies for addressing the challenges of new users, items, and contexts with personalized experiences.
  • Metrics and A/B Testing: Includes considerations for choosing appropriate metrics and performing A/B tests for validating the effectiveness of the system.

Quick Start

Design a recommendation system for a video streaming platform using the provided interview template, considering the two-stage architecture and cold start strategies.

Frequently Asked Questions about recommendation-system

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

FAQPage Schema
How do I design a two-stage recommendation system for a personalized feed?▼

Handle cold start in recommendation systems using strategies for new users, items, and contexts. The framework provides specific approaches to deliver personalized experiences even without historical interaction data.

What metrics should I use for A/B testing a recommender system?▼

Choose metrics for A/B testing a recommender system based on relevance and user engagement. The framework includes considerations for validating system effectiveness through offline training and online serving tests.

How do I mitigate training-serving skew in large-scale machine learning pipelines?▼

Mitigate training-serving skew in machine learning pipelines by utilizing feature stores for efficient data management. This ensures consistent feature transformation between offline training and online serving environments.

Do I need TensorFlow and scikit-learn to build large-scale recommender systems?▼

You need TensorFlow and scikit-learn alongside numpy and pandas to build large-scale recommender systems. These dependencies support the advanced machine learning and data engineering practices required for implementation.

What is the best way to structure a recommendation system design for an interview?▼

The best way to structure a recommendation system design for an interview is using a structured template. It covers two-stage architecture, feature stores, cold start scenarios, and metrics for a comprehensive design challenge response.

Related Skills

Explore More Software Engineering Skills→
andikarachmanandikarachman
14

data-preprocessing

Preprocess raw data into clean, analysis-ready datasets with pandas and numpy.

Community
Advanced
ContextJet-aiContextJet-ai
29

set-up-ab-testing

Split live traffic between control and experimental groups to compare LLM prompt and model outcomes.

Official
Advanced
haJ1thaJ1t

ai-ml-pro

Build, train, deploy, and monitor ML systems with Python and scikit-learn, TensorFlow, or PyTorch.

Community
Advanced
EndUser123EndUser123

chs

Search large Claude Code chat archives with two-stage indexing and filters.

Community
Advanced