Principal Software Engineer, GPU Compute

Overview Roblox is hiring a Principal Software Engineer to lead GPU and AI accelerator compute infrastructure. The role focuses on productionizing, monitoring, and scaling GPU hosts and platforms across data centers and cloud environments. Responsibilities - Lead GPU strategy across Compute, Kubernetes, Machine Bootstrap, Networking, and Cloud teams - Own GPU host lifecycle management, including drivers, firmware, CUDA stack, health telemetry, and remediation of GPU-specific failures - Architect GPU capacity exposure, scheduling, isolation, and Kubernetes integration for GPU and AI workloads - Improve GPU reliability and performance at fleet scale through detection, diagnosis, and automated repair - Evaluate and onboard new GPU/AI accelerator platforms and networking topologies such as NVLink, InfiniBand, and RoCE - Create standards, tooling, and APIs for safe and efficient GPU compute consumption across engineering teams Requirements - 10+ years of experience building and operating large-scale distributed systems and infrastructure - Deep hands-on GPU expertise at the machine management layer and above - Experience operating GPU and AI workloads in production - Strong proficiency in Go or another well-structured programming language - Familiarity with CUDA, GPU scheduling, and high-performance networking - Kubernetes experience for GPU workloads and bare-metal concepts is a plus - History of being the technical anchor for difficult GPU and compute problems Benefits - Annual salary range: $345,040 to $399,420 USD - Eligible for equity compensation - Eligible for benefits - Office-based roles are onsite Tuesday, Wednesday, and Thursday, with optional Monday and Friday presence Source https://careers.roblox.com/jobs/7978492?gh_jid=7978492