Candidate Generation and Ranking Model

Learn about candidate generation and ranking of videos based on user preferences.

We will cover the following

Multi-stage models

Candidate generation model

• Feature engineering
• Training data
• Model

Ranking model

• Feature engineering
• Training data
• Model

1.Multi-stage models

Architecture diagram for the video recommendation system

Video recommendation systems typically use a multi-stage approach for scalability:

Candidate Generation: Filters millions of videos to a smaller, relevant set based on user preferences.

Ranking: Orders the filtered candidates based on how likely the user is to watch them.

This two-step process ensures both speed and accuracy, making it a go-to strategy for companies like Google, Meta, and Netflix.