Performance Test Engineer

📋 Description

• • Own the end-to-end performance strategy for Computer, the AI-native platform that unifies structured and unstructured data across Google Workspace, Jira, Notion, and dozens of other SaaS tools. You will design, build, and run tests that prove our system can ingest, index, and reason over millions of records in real time while keeping sub-second response times for agents and human users.
• • Architect repeatable performance suites that simulate real-world usage patterns—burst ingestion from AirSync connectors, concurrent agentic actions, complex graph traversals in Computer Memory, and high-throughput customer-facing extensions. Your benchmarks will become the single source of truth that product, platform, and customer success teams rely on before every release.
• • Instrument every layer of the stack—from Rust micro-services and vector stores to Kubernetes clusters and edge caches—so that regressions are caught in CI/CD within minutes, not days. You will define SLIs/SLOs for latency, throughput, and resource efficiency, then build dashboards that surface actionable insights to engineering leaders.
• • Partner with distributed teams in the US, India, and the Philippines to embed performance thinking into daily workflows. You will run weekly “perf clinics,” review design docs for scalability risks, and champion a culture where every engineer asks, “How will this behave at 10× scale?”
• • Reproduce and isolate production incidents reported by global enterprise customers. You will create minimal, reproducible test cases, profile hotspots with eBPF, flame graphs, and distributed tracing, then hand off precise fixes to feature teams—turning customer pain into measurable reliability wins.
• • Continuously evolve our test environments. Today we spin up ephemeral, full-stack replicas in AWS and GCP using Terraform and Helm; tomorrow you might add chaos-mesh fault injection or synthetic customer data generators that mirror Fortune-500-scale topologies.
• • Publish quarterly performance reports that go beyond raw numbers—story-telling how optimizations reduced cloud spend by 30 % or cut agent response time from 800 ms to 120 ms. These reports will be read by Khosla Ventures, Mayfield, and our largest customers as proof of DevRev’s technical excellence.
• • Contribute to open-source tools and standards that the broader performance community adopts. Whether you’re upstreaming a patch to k6, writing a white-paper on vector-database benchmarking, or speaking at a meetup, you will amplify DevRev’s reputation as a thought leader in high-scale AI systems.