Assess students' ability to apply Bloom filter principles to reduce disk I/O in large-scale key-value stores, focusing on design trade-offs (memory vs false-positive rate, filter placement, update vs read optimization), parameter tuning (optimal m, k for expected n and target false-positive rate), variant selection (standard, counting, blocked, partitioned), workload effects (read/write ratio, key distribution), implementation concerns (concurrency, persistence, false-positive mitigation), and evaluation metrics (I/O reduction, latency, throughput, memory overhead). Include calculation-based tuning scenarios, comparative analysis, and justified design recommendations for specified workloads and constraints.