RAG Interview Questions, issue 12, Jul 16, 2026
The LSH Recall Paradox
Staff ML Engineer interview at Pinterest, and the interviewer asks:
“Your LSH index is missing 30% of true neighbors. You doubled the number of hash tables and recall barely moved. What’s actually happening?”
Don’t say: “I’d add more tables and tune the number of hash functions.”
Why throwing more memory at missing neighbors quietly balloons your latency instead of fixing your search, and the data-dependent reality that actually rescues your retrieval.
The full answer, with the mechanism and the arithmetic, is for paid subscribers on Substack.