Performance

Performance interviews test whether you can improve user-visible speed and system capacity with evidence—not whether you memorized every webpack flag. Strong answers connect frontend vitals, backend latency, scalability patterns, and measurement discipline to real bottlenecks.

This domain spans frontend performance (Core Web Vitals, bundles, splitting, lazy loading, images, caching, virtualization), backend performance (latency, throughput, profiling, queries, pooling, caching, async bottlenecks), scalability (horizontal scaling, load balancing, queues, rate limits, backpressure), and measurement (synthetic tests, RUM, load testing, benchmark pitfalls).

On interviews: for a slow checkout page, name which Core Web Vital regressed, what bundle or image change you would inspect, which API trace shows queueing, and how you would prove the fix with field data plus a load test.

Common pitfalls: optimizing averages while p95 regresses; caching without invalidation; scaling replicas before fixing N+1 queries; trusting lab scores that ignore real devices and networks.

The trade-off is shipping speed versus measurement rigor, complexity of caching and scaling, and operational cost of capacity.

Checklist:

• Start from user-visible symptoms and percentiles.
• Profile before guessing the bottleneck.
• Pair frontend and backend evidence.
• Use child topics for depth by layer.