Staff Software Engineer, Airbnb

• 9+ years of industry experience, with a track record of building and operating large-scale distributed systems in production
• Led the technical direction of a system or multi-team project end to end: wrote the design, aligned stakeholders, delivered it, and operated it
• Deep experience in at least one telemetry domain at scale: metrics (Prometheus, Grafana Mimir, VictoriaMetrics), logs (OpenSearch, Kafka), or distributed tracing (OpenTelemetry, Jaeger, ClickHouse)
• Strong coding skills in Go or Java/Kotlin, and comfortable working in unfamiliar codebases, including upstream open source
• Experience with cloud-native infrastructure: Kubernetes, AWS, and service mesh (Istio, Envoy)
• Clear written and verbal communication, including design docs and tradeoff analysis for engineering leadership

Desirable

• The Observability team builds the platform Airbnb engineers use to understand how their systems behave in production. We collect, process, store, and query metrics, logs, events, and traces from both backend services and client apps. Our metrics platform alone ingests tens of millions of samples per second across more than a billion active time series. AI agents are now consumers of this data alongside engineers: they query telemetry to write and validate code and to help debug production issues. We continue to invest in reliable and efficient real time system for both AI and human access
• As a Staff engineer, you will set the technical direction for a major area of the Observability platform and lead it from design through production. Problems in front of us include:
• Scaling the metrics pipeline (OpenTelemetry, vmagent, Grafana Mimir) while controlling cost through aggregation, sampling, and blocking unused metrics
• Replacing legacy log aggregation with an OpenTelemetry-based pipeline, and evolving log and trace storage on OpenSearch
• Building a fast, reliable query layer for both engineers and AI agents, with rate limiting, access control, isolation, and caching so agent traffic does not slow down incident response
• Giving teams self-service visibility and control over their telemetry cost: unit cost per team and service, and controls for granularity, sampling, and retention
• Improving time to detect with anomaly detection and automated correlation across signals
• Lead the design and delivery of distributed systems that handle tens of millions of events per second, from prototype to production operation
• Write well-crafted, well-tested, readable, maintainable code, and contribute back to the open-source projects we run on
• Align roadmaps across Observability and partner infrastructure and product teams, and drive build, buy, and open-source decisions
• Explain tradeoffs and return on investment to engineering leadership in terms of reliability, cost, and developer productivity
• Mentor engineers on the team and raise the bar on design reviews, operational readiness, and on-call
• Help Airbnb engineers instrument their services and self-serve when debugging production issues