What makes code search useful to an autonomous coding system?

Example Answer

The way I'd think about it is this: Useful code search helps the agent find the implementation path, not just matching strings. That means surfacing the right files, call sites, tests, docs, and adjacent patterns so the agent can understand how a change should fit the repo.

I want search to support both precise lookup and relationship discovery. A good result tells the agent not only where the code is, but how similar problems are solved elsewhere and what modules or tests might be impacted.

Search that returns lots of lexical matches without structural relevance is often worse than no search because it wastes the agent’s limited attention budget.

What matters in an interview is not only knowing the definition, but being able to connect it back to how it changes modeling, evaluation, or deployment decisions in practice.

Common Poor Answer

A weak answer is saying code search is useful if it returns all possible matches quickly. Relevance and structural signal matter more than raw volume.