Performance Engineer (6 Month Remote Contract)

📋 Description

• • Own the end-to-end performance lifecycle for Fortune-500-grade platforms that serve millions of transactions daily. You will architect, script, execute, and interpret load, stress, soak, and spike tests that mirror real-world customer behavior across distributed microservices, Kubernetes clusters, and multi-region AWS topologies.
• • Partner shoulder-to-shoulder with Engineering, QE, Product, DevOps, and Support to translate vague performance SLAs into precise, measurable acceptance criteria. You will embed early in the SDLC to influence design decisions, review code for scalability anti-patterns, and ensure every new feature ships with quantifiable performance guardrails.
• • Build and evolve a next-generation performance framework using k6 or Gatling written in JavaScript/TypeScript. Your code will become the golden standard that other QE teams fork, extend, and reuse—complete with reusable modules, CI hooks, and self-service dashboards.
• • Instrument, collect, and analyze telemetry from DataDog, CloudWatch, and custom exporters to create mathematically defensible models that predict response-time percentiles, throughput ceilings, and cost-per-transaction under varying load shapes.
• • Hunt bottlenecks wherever they hide—JVM GC pauses, mis-tuned thread pools, noisy-neighbor pods, under-provisioned EBS IOPS, or API gateway rate limits—and deliver concise, data-backed remediation playbooks to squads that own the code.
• • Champion capacity-planning rituals that translate marketing forecasts into infrastructure budgets. You will simulate Black-Friday-level traffic six months in advance and negotiate reserved-instance purchases that save six-figure cloud spend.
• • Mentor and pair with junior performance testers across multiple concurrent projects, teaching them how to isolate variables, design statistically significant experiments, and present findings that compel action from skeptical stakeholders.
• • Continuously refine our definition of “customer-like” load: replaying production traffic captures, anonymizing PII, and replaying it at 10× scale to uncover edge-case memory leaks before customers feel them.
• • Document every test plan, result, and recommendation in living Confluence pages that become institutional memory. Your diagrams and runbooks will be referenced by on-call engineers at 3 a.m. when alerts fire.
• • Propose and drive process improvements—whether that’s shifting performance gates left in the pipeline, introducing chaos-engineering experiments, or lobbying for new observability budgets—that reduce mean-time-to-resolution by measurable percentages.