Principal Software Engineer, Performance

About the Role

Airbnb's infrastructure is scaling fast in the AI era: more compute, more GPU spend, more complexity across the stack. The Infrastructure Performance team exists to make sure that growth doesn't come at the cost of speed, reliability, or efficiency. We treat performance as a first-class engineering property - measurable, continuously improving, and enforced across the software lifecycle.

This is the most senior performance engineering role at the company. In this role, you won't be handed a roadmap. You'll build one.

The Performance team owns performance across the full stack: CPU and GPU fleet utilization, backend service efficiency, end-to-end request latency, storage, ML infrastructure, and observability. That means instrumenting and profiling at scale, discovering hot paths, shaping how teams across the company think about performance SLOs, and influencing build vs. buy decisions on emerging hardware and software.

The team also builds the tooling that lets every backend team self-serve performance insights, scaling good decisions beyond your direct involvement.

On the product side, as the TL for Performance team, you will partner with engineering teams to tie infrastructure performance directly to user experience and business outcomes. When a host's listing loads faster or a guest's search returns cleaner results, some of that is the win enabled by the Performance team.

Why This Role Is Unusual

Most performance engineering roles are reactive: triage a regression, close the ticket. This one is different. You'll be setting the strategy, defining what "good" looks like, and building systems that make performance a default property of how Airbnb ships software. That requires someone who can operate at both the systems level and the organizational one.

Responsibilities

• Define and drive Airbnb's long-term performance strategy, spanning product performance, infrastructure efficiency, and business objectives for scale and growth
• Lead the architecture and development of performance profiling and instrumentation infrastructure, covering CPU, GPU, memory, request hot paths, utilization, and deployment events, making these capabilities available to all backend teams
• Partner with infrastructure teams across compute, reliability, backend frameworks, and AI Infra to ensure the fleet operates at optimal utilization
• Connect performance outcomes to business objectives and company-wide SLOs, and guide engineering teams in keeping the stack scalable and efficient
• Evaluate emerging hardware and software technologies, engage with the external solutions ecosystem, and advise on build vs. buy decisions in areas of strategic importance
• Mentor and technically uplevel engineers across the organization through design reviews, architectural guidance, and performance best practices

About You

• 12+ years of performance engineering experience in high-scale, high-growth production environments
• Deep understanding of how software and hardware systems interact at scale, including architectural patterns for performance-critical stacks. You know the difference between a bad algorithm and a bad cache eviction policy, and you're comfortable in the kernel, in the JVM, and in a position to drive change across the stack