Senior Software Engineer (Traffic Proxy)
Posted Updated
Description
- The Traffic Proxy Services team powers the front line of Roblox — operating the global edge network and high-performance Layer 7 proxies that handle billions of web and API requests daily while defending the platform against large-scale automated attacks
- In this role, you’ll help pioneer the next generation of our edge infrastructure: replacing legacy proxy configurations with programmable, memory-safe Rust services running on Kubernetes, and building intelligent, automated defense systems
- Working alongside principal engineers and key partners across Security, Reliability, and Platform teams, you’ll tackle high-stakes systems challenges where efficiency, scale, and uptime are critical
- Design and build high-performance services that sit directly in the request data path, where performance, reliability, and security are of utmost priority
- Effectively leverage AI tools to speed up research, development, and troubleshooting while keeping a critical eye
- Evolve our edge architecture toward something simpler and programmable, migrating policy incrementally with parity testing and exercised rollback
- Improve traffic classification and intelligence to detect and mitigate DDoS and abuse traffic
- Build feedback loops that automate mitigation and reduce MTTR
- Raise the efficiency of the edge
- Own your work in production as part of the team’s on-call rotation, and raise the reliability bar through design, code review, incident review
- Collaborate across Networking, Security, Reliability, and Compute to deliver on Infrastructure and company priorities
- Be a gatekeeper for a high code quality bar, architectural designs, and pragmatic vs. long-term approaches
- Uphold our company vision by putting the Roblox community first and exemplifying our values
- Able to write a design doc that survives review by senior stakeholders across several teams, and to disagree productively when it doesn’t
- Motivated by problems where the blast radius is real and the feedback loop is immediate
- Able to independently design, build, and operate systems on Kubernetes end to end, and to reason about node-level behavior when latency and resource isolation matter
- Deeply comfortable with networking fundamentals: TCP, TLS, HTTP/1.1 and HTTP/2, QUIC and the internals of at least one production L7 proxy (HAProxy, Envoy, NGINX, Pingora or similar). Able to troubleshoot a live network problem
- Familiar with how the kernel and userspace interact, and comfortable tuning the kernel to optimize for efficiency and reliability
- An experienced systems engineer with 7+ years building and operating backend or infrastructure software, including time owning a tier-0 service in production
- Experienced with high-performance packet processing in and around the kernel — kernel-bypass and kernel-path techniques such as eBPF/XDP
- Genuinely curious. You pull on the thread until you understand why a system behaves the way it does, and you would rather learn an unfamiliar layer of the stack than route around it
- Fluent in a systems language. Rust strongly preferred; strong Go experience with the appetite to work in Rust is welcome