179 questions
No questions match those filters.
← All questions
Retrieval & EmbeddingsHard· HNSW· Memory Optimization· Vector Databases· Infrastructure Scaling
Your team’s uncompressed HNSW cluster for this corpus n...
This is one of the questions in the full AI/ML interview bank. Pro unlocks all 1789 questions; Premium includes the same bank plus the highest daily Practice limit.
See plansThe core issue is separating the axes of corpus size and query load. Because the corpus size is unchanged, the memory-driven shard count cannot be reduced without compromising the HNSW index performance. The current node count is dictated by the need to handle the full system rate under scatter-gather retrieval. Therefore, any additional nodes added would be for replicas to handle throughput, not for sharding the corpus itself.