Compute Platform Lead: Scale Multi-Cloud GPU Clusters

Oscar Faye; New York (NY); On-site; USD 250,000 - 360,000; Full time

Benefits offered by this job

Health cover

Parental leave

Flexible time off

Visa sponsorship

Equity

Job summary

Oscar Faye is seeking a Compute Platform Lead to own the GPU compute layer that powers frontier-scale model training. You will lead a Kubernetes-based platform spanning multiple cloud providers, tackling scheduling, health, and performance challenges at scale while staying hands-on to ship code.

You will build a small, highly capable team, shape architecture for auto-remediation and topology-aware scheduling, negotiate major compute deals, and guide the lab toward larger, petabyte-scale storage

Qualifications

• Proven track record of building and leading infrastructure or systems teams while staying hands-on.
• Deep systems engineering experience with how whole clusters behave and how to maintain them.
• Strong coding ability and the credibility to earn a senior team's technical trust.
• Depth in at least one of orchestration, storage or GPU hardware; GPU knowledge beyond standard Kubernetes (e.g. NCCL) is a plus.
• Experience with high-performance storage across data centres and with checkpointing at scale is a plus.
• Commercial judgement managing vendors and closing significant deals.

Responsibilities

• Build, mentor and grow a team of strong systems engineers, scaling towards about 10.
• Own fleet reliability and availability: scheduling across clouds, cluster management, and getting ready for next-generation GPUs and much larger clusters.
• Make the architecture calls on automatic remediation, topology-aware scheduling, capacity planning, hardware debugging, and fleet-wide monitoring and benchmarking.
• Work with the training teams to design fault tolerance, node health checks and remediation together.
• Own vendor relationships, including negotiating and running major compute deals.
• Over time, take on multi-cloud storage, petabyte-scale data replication and GPU-to-GPU network performance.

Skills

Infrastructure leadership

Systems engineering

Coding ability

Orchestration / GPU hardware

Vendor management

Job description

Oscar Faye is seeking a Compute Platform Lead to own the GPU compute layer that powers frontier-scale model training. You will lead a Kubernetes-based platform spanning multiple cloud providers, tackling scheduling, health, and performance challenges at scale while staying hands-on to ship code.

You will build a small, highly capable team, shape architecture for auto-remediation and topology-aware scheduling, negotiate major compute deals, and guide the lab toward larger, petabyte-scale storage