About the Role:AI models reshaping how our community creates, plays, and connects, all run on Compute Platform. As
Senior Product Manager, Compute Platform, you'll set the strategy and roadmap for Roblox's next-generation AI infrastructure: the rapidly growing fleet of GPUs and AI accelerators spanning Roblox core and edge data centers, and public cloud that decides how fast we can train, serve, and scale every model on the platform. You'll own the products that turn raw GPU hosts into reliable, production-ready AI compute - driver and firmware management, fleet-wide health and performance, and the abstractions product teams across Roblox build on.
You Will:- Drive strategy and roadmap for Compute Platform spanning Managed Kubernetes (Roblox Kubernetes Service), Managed Compute Services and other critical distributed systems, and our fleet of GPU and CPU machines managed via unified Fleet APIs - all across on-prem and cloud.
- Drive the evolution of our Compute infrastructure to support Roblox's most critical workloads - from AI to Storage to Data Analytics and more - each with their own distinct requirements.
- Build and scale our GPU infrastructure to support training and inference for frontier models, enabling innovation at scale while optimizing cost-to-serve.
- Stay laser-focused on Compute Platform Reliability for AI and other Roblox workloads, lowering mean-time-to-detection and recovery from failures.
- Partner closely with seven key platform teams to understand their use cases and empower them with Compute Platform primitives enabling them to build on top easily.
- Lead cross-functionally, acting as the connective tissue between Compute Platform Users, Engineers, AI researchers, Product, Finance, and Roblox leadership.
You Have:- 7+ years of product management experience, focused on Compute infrastructure or distributed systems at scale.
- Deep practical understanding of Kubernetes internals and control plane components (API Server bottlenecks, Etcd scaling limitations, Kubelet behavior, etc) and experience productizing custom Kubernetes Operators, Controllers, and Custom Resource Definitions (CRDs) to extend platform capabilities beyond vanilla implementations.
- Deep familiarity with GPU/accelerator architecture and scheduling challenges, including topology-aware placement, preemption, hardware affinity constraints, and mechanisms to validate resource readiness before scheduling to avoid wasted compute cycles.
- Familiarity with cloud-native service networking including microservices, CNIs, and enterprise-grade service mesh architectures.
- Built production-grade compute platforms where efficiency, reliability, and developer experience were designed-in from day one.
- The ability to balance the needs of multiple users and stakeholders and make optimal tradeoffs between utilization, latency, cost and time-to-ship while maintaining reliability.
- A builder mindset - you are passionate about prototyping, evolving products through rapid iteration, and leveraging AI for ideation and unlocking value for users.
- Experience building Compute infrastructure on AWS, Google Cloud Platform (GCP), Azure or other cloud providers. (preferred)
- Background in AI model development, training, inference. (preferred)
- Kernel-level experience or familiarity with custom kernel drivers. (preferred)
- Experience building agentic systems for Compute or infrastructure. (preferred)
For roles that are based at our headquarters in San Mateo, CA: The starting base pay for this position is as shown below. The actual base pay is dependent upon a variety of job-related factors such as professional background, training, work experience, location, business needs and market demand. Therefore, in some circumstances, the actual salary could fall outside of this expected range. This pay range is subject to change and may be modified in the future. All full-time employees are also eligible for equity compensation and for benefits as described on
this page.
Annual Salary Range
$280,540-$330,950 USD
Roles that are based in an office are onsite Tuesday, Wednesday, and Thursday, with optional presence on Monday and Friday (unless otherwise noted).
