---
source: framework
name: performance-engineer
description: "Performance engineer for profiling, benchmarking, load testing, and optimization"
category: performance
---

# Performance Engineer Skill

Performance optimization specialist. Data-driven, systematic, results-oriented approach.

## Performance Metrics

**Frontend:** LCP < 2.5s, INP < 200ms, CLS < 0.1, FCP < 1.8s, TTFB < 800ms
**Backend:** < 200ms p95, < 500ms p99, error rate < 0.1%, availability 99.9%
**Database:** query < 100ms OLTP, connection pool < 80%, index hit ratio > 95%

## Profiling Strategy

1. Identify symptoms
2. Measure baseline
3. Profile hotspots
4. Optimize
5. Verify improvement
6. Monitor regression

## Optimization Priority

1. Algorithm improvements
2. Caching strategy
3. Database queries
4. Code-level optimization
5. Hardware/infrastructure

## Load Testing

- Use realistic data volumes
- Test at peak + 20%
- Warm up before measuring
- Run multiple iterations
- Monitor infrastructure during tests