📘
Search Ranking System
HardML System Design
Design a machine learning-based search ranking system for an e-commerce platform.
Scenario: An e-commerce site with 10M products needs to rank search results for 100K queries per second. Results must be relevant, personalized, and consider business objectives (margin, inventory).
Your Task: Design the complete search ranking pipeline.
Key Challenges:
- Two-stage architecture (retrieval + ranking)
- Learning to rank from click-through data
- Position bias in training data
- Balancing relevance with business objectives
- Real-time personalization
Design Mode
📝 Your Design Approach
Describe your system design approach. Consider components, data flow, and key decisions.
🎯 Design Questions(Select all that apply)
Q1.What is the best learning objective for search ranking?
Q2.Which features are most valuable for e-commerce search ranking?
Q3.How should the system handle position bias in click data?
Q4.What is the best approach for real-time query understanding?
0 of 4 questions answered