Everything in Databases & Architecture
NVIDIA's AVO harness ran Claude Opus 5 autonomously for seven days straight
NVIDIA paired Claude Opus 5 with a new harness called AVO to run long-horizon autonomous tasks, including a seven-day GPU kernel optimisation run and a separate reasoning benchmark. AVO uses persistent memory and a supervisor process so the agent keeps working past a single context window instead of restarting.
AI code volume is pushing code review earlier, ahead of the pull request
AI now produces more code than humans can realistically review, with Meta's lines of code per human-landed diff reportedly up 106% in a year. One response argues review should move earlier, into pairing and design sessions, leaving the pull request to catch formatting and known security issues.
Diagnosing high Redis eviction and cache misses
Use INFO memory and stats to confirm pressure, check fragmentation ratio, pick LFU over LRU for skewed access, set sane TTLs.
Modeling IoT data with tags and fields
Tags are indexed identifying metadata, fields are unindexed measured values, and tag cardinality drives memory.
Problems the Lakehouse architecture solves
Lakehouse adds ACID transactions, schema enforcement, and time travel on cheap object storage.
Why choose Kafka over a REST endpoint for ingestion
Kafka buffers spikes, decouples producers from consumers, replays and fans out durably.
Cutting managed database costs without breaking SLOs
Pool connections, prune and tune indexes, offload reads, tier or partition cold data, right-size storage IOPS.
Designing an HA/DR strategy for an OLTP database
Sync standby in-region for zero data loss, async cross-region for DR, automated failover with a quorum.
Diagnosing and optimizing a slow production query
Read the EXPLAIN ANALYZE plan, find the costly node, then fix via indexing, rewrite, or stats.
Optimizing queries on a billion-row fact table
Partition to prune scans, index for selective lookups, materialize views to precompute aggregates; each adds write or maintenance cost.
Logical vs physical query plans and the optimizer
Logical plan says what (relational algebra, no algorithms); physical plan says how (specific operators); cost-based optimizer enumerates physical options and picks the cheapest using statistics.
How databases implement GROUP BY aggregation
Hash aggregation builds a hash table keyed by group holding running aggregates; sort aggregation orders rows then aggregates adjacent groups; optimizer picks based on data and memory.
The Volcano iterator model of query execution
Each operator exposes open/next/close, parents pull tuples from children, uniform composable interface, pipelined low memory.
Phantom reads and how serializable prevents them
New rows matching a predicate appear between reads; classic Repeatable Read locks existing rows not ranges; Serializable uses range or predicate locks.
Choosing a time-series database for metrics
High-ingest timestamped writes, time-window queries, retention and downsampling, time-optimized compression.
Diagnosing database latency layer by layer
Split total time into pool-wait, query execution, and ORM-generated query patterns; use metrics at each layer.
Sessionizing clickstream events into sessions
Order events per user, split on inactivity gap, assign session ids, pick event or session grain.
Vectorized query execution and its speedups
Process column batches per operator call, amortize per-tuple overhead, use cache locality and SIMD.
The buffer pool's role in database IO
Caches pages, serves reads from RAM, buffers dirty writes flushed later, uses eviction like LRU.
Predicate pushdown and why it speeds queries
Apply WHERE conditions at the scan or remote source, prune partitions and rows early, shrink data movement.
We are hiring for this. Every open role lists the topics its interview covers, so you can prepare for the real thing rather than guessing.
See open roles