Twitch rebuilt its search system from scratch starting in 2019, replacing a third-party dependency with an in-house system that uses Amazon OpenSearch for retrieval and machine learning models to re-rank search results for relevance.

The search system has four components: an API handling experimentation and orchestration, a retrieval system built on Amazon OpenSearch returning lexical matches across live streams, channels, categories, and videos, a ML-powered ranking system that re-ranks top results, and training pipelines to continuously train and deploy ranking models. The Search team sits within the Community Discovery Machine Learning organization. Twitch's 2021 Research Fellowship listed "Information Retrieval" — including "search and ranking, document categorization and clustering" — as an approved research area.