GopherCon 2025: To Trie or Not to Trie: Typeahead Completion Using Redis - Garrett Denis

Why watch

This talk demonstrates how to build a high-performance typeahead completion system by implementing a Redis-backed trie to replace inefficient Postgres LIKE queries for @mentions. Garrett Denis shares the complete migration strategy, including the data structure choices, implementation details, and how they achieved 80% latency reduction under heavy load. Ideal for Go developers building search or autocomplete features who want to learn practical Redis optimization techniques.