Lukas Reinhart Senior Performance Engineer
Seattle, WA • lukas.reinhart@gmail.com • +1 (206) 555-0187
Profile Summary
- Senior Performance Engineer with 7+ years specializing in load testing, application profiling, and capacity planning for SaaS scaleups, payment platforms, and high-traffic web properties.
- Hands-on across the performance stack with load and stress testing (k6, JMeter, Gatling, Locust), application profiling (async-profiler, py-spy, pprof, Chrome DevTools), database tuning (PostgreSQL, MySQL, Redis), and APM/observability (Datadog, New Relic, Grafana, OpenTelemetry).
- Strong grounding in SLO-driven engineering, workload modeling against production traffic, p99 latency budgeting, capacity forecasting, and cloud cost-performance tradeoffs. Comfortable embedding continuous performance testing in CI/CD pipelines in performance and reliability engineering at scale.
- Comfortable working with backend, SRE, platform, and product teams to define performance requirements, build representative test environments, run pre-launch hardening, and own post-incident performance reviews.
- Bias toward measurable performance over vanity metrics, mentor to engineers learning to read flame graphs and APM dashboards, and frequent contributor to internal performance playbooks and reliability forums.
Technical Skills
- Languages:
- Java, Python, Go, JavaScript/TypeScript, SQL
- Load & Stress Testing:
- k6, JMeter, Gatling, Locust, NeoLoad, Artillery
- Profiling & APM:
- Datadog, New Relic, Dynatrace, async-profiler, py-spy, pprof, Chrome DevTools, Grafana, OpenTelemetry
- Frontend Performance:
- Lighthouse, WebPageTest, Core Web Vitals, RUM, bundle analysis, CDN tuning
- Databases:
- PostgreSQL, MySQL, Redis, MongoDB, query plan analysis, indexing, connection pooling
- Infrastructure & Cloud:
- AWS EC2/EKS/RDS/CloudFront, autoscaling, HPA, load balancing
- Methodology:
- SLO/SLI design, workload modeling, p50/p95/p99 budgeting, capacity planning, continuous performance testing, profiling-driven optimization
- CI/CD & Tooling:
- GitHub Actions, Jenkins, GitLab CI, Docker, Kubernetes, Helm
Education
University of
Washington
B.S. in Computer
Engineering
Work Experience
Twilio
Senior Performance
Engineer
- Owned the performance engineering function supporting 350+ microservices and 2 billion daily API requests across the Programmable Messaging and Voice platforms, leading end-to-end design across performance budgets, load-test infrastructure, and capacity headroom.
- Partnered with SRE and product teams to define measurable SLOs for API gateway and webhook delivery services, codifying p99 < 220ms and 99.95% availability targets across 18 product surfaces and baking them into the continuous performance testing pipeline.
- Built production-mirrored workload models from 6 months of RUM and access-log data, applying per-tenant request-mix and Poisson-arrival patterns to surface scaling cliffs at 3.2x peak traffic before each major release.
- Profiled hot paths with async-profiler and JFR across the JVM-based messaging dispatcher, identifying a GC-pressure regression that brought p99 from 410ms to ~160ms after applying G1 tuning and allocation reductions.
- Tuned PostgreSQL and Redis read paths for the delivery-status service, partnering with the DBA team on index redesign and connection-pool sizing to cut slow-query p95 from 1.8s to ~280ms across 40 query patterns.
- Right-sized EKS node groups and EC2 autoscaling for the inbound webhook pipeline, applying per-pod request budgeting and HPA tuning to lift steady-state utilization from 35% to ~62% and save ~$410K per year in compute spend.
- Built Datadog and Grafana performance dashboards tracking p50/p95/p99 latency, error rate, and saturation, partnering with 9 service teams to catch ~78% of perf regressions in pre-prod and lift release-time MTTD from 2 hours to ~12 minutes.
Expedia
Performance
Engineer
- Owned client-side performance for the Expedia booking-flow web app, applying bundle-splitting, image lazy-loading, and edge caching via CDN with Lighthouse and WebPageTest to lift LCP from 4.2s to ~1.9s and Core Web Vitals good-rate from 38% to ~78%.
- Built load and soak test suites with JMeter and k6 for the search and pricing services, applying realistic user-journey traces to expose memory leaks that cut steady-state heap usage from 3.1GB to ~1.6GB.
- Profiled the hotel-search Node.js backend with clinic.js and 0x flame graphs, refactoring event-loop bottlenecks and async patterns to bring server-side TTFB from 480ms to ~190ms across 12 endpoints.
- Modeled peak-season capacity for the booking and pricing platforms using headroom and queue-length analysis, right-sizing AWS reserved capacity to absorb 3.5x summer-spike traffic without incidents.