MTS on ClusterMAX (GPU Infrastructure)

Location: United States, New York, Mexico, San Francisco, or Canada; in-office or remote. Office hours.

Company

hirify.global is an independent research and analysis firm focused on the semiconductor and AI industries, covering the supply chain from fabrication and AI models to CUDA kernels and GPU cloud infrastructure.

What you will do

• Lead development of ClusterMAX benchmarks for storage I/O and bandwidth, NCCL/RCCL collectives, fault tolerance, goodput, and multi-tenant security.
• Deploy and evaluate GPU clusters across hyperscalers and neocloud providers, including GB300, GB200, B300, B200, H200, MI355X, and TPUv7 systems.
• Extend TCO and goodput methodologies into automated, reproducible tests.
• Work with executives and engineers at neoclouds, hyperscalers, and chip vendors.
• Author technical research analyzing benchmark results, reliability, security, and ease of use.

Requirements

• Hands-on experience operating GPU clusters with Slurm and/or Kubernetes.
• Experience with InfiniBand or RoCE fabrics and distributed storage.
• Strong Python and shell scripting skills, including benchmark harnesses and CI pipelines.
• Understanding of distributed training and inference failure modes, including node failures, checkpointing, blast radius, and recovery.
• Security mindset focused on multi-tenant isolation and bare-metal versus virtualized infrastructure.

Nice to have

• Experience with technical due diligence, TCO analysis, and client-facing technical communication.
• Experience similar to a technical consultant role.

Culture & Benefits

• Autonomous, research-driven environment with direct exposure to AI and semiconductor industry discussions.
• Generous paid time off.
• Office stipends and competitive medical, dental, and vision healthcare.
• Support for conferences and ongoing learning.
• Employment eligibility is verified through E-Verify.