Compute Platform Lead: Multi-Cloud Kubernetes & GPU

Socket.dev; San Francisco (CA); On-site; USD 210,000 - 320,000; Full time

Benefits offered by this job

Top-tier compensation and equity

Stock options

Health & wellness

Meals provided

Paid parental leave

Unlimited PTO (US)

Sponsorship support

Team off-sites

Job summary

Reflection is seeking a Compute Platform Lead to lead a Kubernetes-based compute layer spanning multi-cloud environments. You will mentor a team of systems engineers, guide architectural decisions, and stay hands-on to contribute as an IC on critical path work.

You will manage day-to-day execution, collaborate with training teams, and drive next-generation GPU deployments and larger clusters. The role demands strategic thinking and strong cross-team communication across a fast-paced research

Qualifications

• Experience building, mentoring, and growing systems or infrastructure teams while staying technically hands-on.
• Deep systems-level engineering with focus on cluster-wide behavior and maintenance.
• Strong coding ability and the credibility to earn the technical trust of a strong team.
• Depth in at least one of orchestration, storage, or GPU hardware — with the ability to learn the rest.
• Alignment with a Kubernetes-first architecture.
• Cloud storage expertise — managing high-performance data products across multiple data centers is a plus.
• Experience managing vendors, including negotiating and operating important deals.
• Ability to guide strategy and drive execution across a multi-cloud, large-fleet environment.

Responsibilities

• Build, mentor, and grow a high-performing team of systems engineers. Coach and support your reports in understanding, and pursuing, their professional growth.
• Provide front-line leadership of engineering efforts to keep the compute fleet reliable and highly available — multi-cloud scheduling, cluster management, and the path to next-generation GPUs and increasingly larger cluster sizes.
• Stay hands-on: become familiar with the team's technical stack enough to make targeted contributions as an individual try.
• Manage day-to-day execution: prioritize the team's work and manage projects in a highly dynamic, fast-paced environment.
• Guide technical and architectural decisions, emphasizing scalability, robustness, and reliability — automatic remediation, topology-aware scheduling, capacity planning, rapid hardware debugging, and cluster-wide monitoring and performance benchmarking.
• Work closely with our training teams to co-design fault tolerance, node health checks, and remediation, and manage the vendor relationships and important deals the compute fleet depends on.
• Prepare the fleet for what's next: next-generation GPUs and larger clusters, and — longer term — multi-cloud storage, petabyte-scale data replication, and GPU-to-GPU network performance.
• Raise the bar for technical judgment, prioritization, communication, and execution in a fast-moving environment.

Skills

Leadership

Systems engineering

Kubernetes

GPU hardware

Multi-cloud

Vendor management

Communication

Architectural decisions

NCCL

Tools

NCCL

Job description

Reflection is seeking a Compute Platform Lead to lead a Kubernetes-based compute layer spanning multi-cloud environments. You will mentor a team of systems engineers, guide architectural decisions, and stay hands-on to contribute as an IC on critical path work.

You will manage day-to-day execution, collaborate with training teams, and drive next-generation GPU deployments and larger clusters. The role demands strategic thinking and strong cross-team communication across a fast-paced research