Kaushik Rangadurai
Recommendation systems researcher at Meta
Kaushik Rangadurai works on recommendation systems at Meta, with a focus on LLM-based recommendations, retrieval and graph representation learning. He co-authored the GenRecs Tokenization publication and has presented on tokens for recommender systems.
Explore Kaushik’s work
Shows, ideas and places to learn from Kaushik Rangadurai.
Browse by format. Each card shows the recorded connection.
Channels & websites1
Newsfeed1
Meta explains how GenRecs Tokenization helps ad recommendations learn across items
Meta published an engineering explanation of GenRecs Tokenization on 19 October. The work describes a more stable vocabulary for representing items in ad recommendations, helping models share information across related content instead of relying only on identifiers with little meaning outside a particular item.
Image: MetaThis is a collection of published credits and official links, not an exhaustive CV or a points total. A guest appearance is different from hosting a show. Community and event links identify the named organiser; they do not imply membership or a speaking appearance.