Staff Engineer, Compute Platform & GPU Infra

Visa Hunt; New York (NY); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Top-tier compensation

Stock options

Health & wellness

Meals provided

Parental leave

Unlimited vacation

Visa sponsorship

Team events

Job summary

Reflection is a research lab focused on making intelligence open and accessible. The Compute Platform team runs a Kubernetes-based platform across multiple neo-clouds, ensuring healthy compute, fault tolerance, and high availability for large GPU workloads.

You will co-design remediation strategies with training teams and drive platform evolution into multi-cloud storage and petabyte-scale data replication.

Qualifications

• Systems-level engineering with a focus on cluster-wide behavior and maintenance.
• Strong coding ability and a demonstrated focus on systems or GPU infrastructure.
• Deep GPU hardware knowledge beyond standard Kubernetes, e.g., familiarity with NCCL.
• Alignment with a K8s-first architecture.
• Cloud storage expertise across multiple data centers, managing high-performance data products and checkpointing at scale.

Responsibilities

• Cluster Management: Build and maintain tools for automatic remediation, topology-aware scheduling, capacity planning and rapid hardware debugging.
• Platform Engineering: Design and iterate on our cluster management stack for workloads across large, multi-GPU fleets.
• Monitoring & Observability: Implement comprehensive cluster-wide monitoring focusing on durability and performance benchmarking.
• Roadmap Execution: Prepare infrastructure for next-generation GPU deployments and larger multi-cloud storage and network performance goals.

Skills

Systems-level engineering

GPU infrastructure

Kubernetes

Cloud scaling

Tools

NCCL

Job description

Reflection is a research lab focused on making intelligence open and accessible. The Compute Platform team runs a Kubernetes-based platform across multiple neo-clouds, ensuring healthy compute, fault tolerance, and high availability for large GPU workloads.

You will co-design remediation strategies with training teams and drive platform evolution into multi-cloud storage and petabyte-scale data replication.