ASTs for Search: Balancing Complexity and Performance in Boolean Query Engines

The new way GitHub Issues handles search, using ASTs for those nested queries and boolean operators, really makes you think about design choices. ASTs are awesome for complex search rules, but I gotta wonder how they keep things fast, especially when there are tons of users. How do teams usually turn a tricky AST into a super-efficient Elasticsearch query, or something similar, without making everything slow? Are there standard ways to optimize ASTs, or is it mostly about what the search engine itself can do? I'm picturing a crazy user query just bogging down the whole system for everyone.