performance-optimizer

---

name: Performance Optimizer description: Identify and fix performance bottlenecks in application code.

Performance Optimizer Skill

Systematically identify and resolve performance issues to improve latency, throughput, and resource usage.

Profiling Approach

• Identify hot paths: database queries, external API calls, CPU-intensive computations, and I/O operations.
• Use language-specific profilers (cProfile, Node.js --inspect, Go pprof) to gather baseline metrics.
• Focus on the 80/20 rule: optimize the small fraction of code causing most of the delay.

Caching Strategies

• Application cache: In-memory (Redis, Memcached) for frequently accessed, rarely changing data.
• HTTP cache: ETag, Cache-Control headers for static and semi-static responses.
• Query cache: Cache expensive DB queries with appropriate TTL and invalidation.
• Computation cache: Memoize pure functions with deterministic inputs.
• Document cache invalidation rules to avoid stale data.

Query Optimization

• Eliminate N+1 queries: use batch loading, JOINs, or eager loading.
• Add indexes for filter, sort, and join columns. Avoid over-indexing write-heavy tables.
• Use EXPLAIN/EXPLAIN ANALYZE to verify query plans. Optimize slow queries before adding indexes.
• Consider read replicas or connection pooling for high concurrency.

Lazy Loading and Deferral

• Defer non-critical work: background jobs, async processing, and queue-based architectures.
• Lazy-load large resources (images, modules) when needed. Use code splitting in frontend bundles.
• Stream large responses instead of buffering in memory.

Output

Provide concrete code changes with before/after metrics where possible. Prioritize high-impact, low-risk optimizations.