У роли четкий фокус на AI-инфраструктуре: распределенное обучение, производительность GPU и рабочие кластеры. Полностью офисный формат и требование разрешения на работу в США сужают круг тех, кому подойдет вакансия.
Только офис
Нужно разрешение на работу
Актуальный стек
Перспективное направление
Понятные задачи
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
ML Performance Engineer (AI): Validating AI compute clusters through large-scale training and inference workloads with an accent on distributed ML performance, GPU profiling, and platform reliability. Focus on diagnosing collective communication issues, benchmarking AMD Instinct GPU workloads, and building repeatable Kubernetes and Slurm validation workflows.
Location: Las Vegas, Nevada, United States; onsite
Company
TensorWave delivers a secure, reliable, resilient cloud platform for AI compute at scale.
What you will do
Validate new cluster deployments through large-scale training and inference workloads before customer handoff.
Identify, reproduce, and diagnose performance and reliability issues with infrastructure and platform teams.
Use managed Kubernetes and Slurm platforms from a customer workload perspective and provide improvement feedback.
Run internal and external model benchmarks across training and inference.
Build repeatable tooling and workflows for cluster validation and benchmarking.
Support customer clusters and workloads while tracking distributed training and inference performance on AMD Instinct GPUs.
Requirements
Bachelor of Science in Computer Science, Computer Engineering, or a related technical field, or equivalent practical experience.
3+ years of experience running ML workloads on GPU clusters.
Hands-on distributed training experience with frameworks such as PyTorch FSDP, Megatron-LM, DeepSpeed, or torchtitan.
Working knowledge of Slurm and Kubernetes from a user and workload perspective.
Experience profiling and optimizing GPU workloads, including collective communication performance.
Strong Python skills and solid Linux fundamentals.
Clear written communication of performance findings to engineering and non-technical audiences.
Employment requires authorization to work in the United States; offers are contingent on identity and work-authorization verification.
Employment may be contingent on a job-related background check where permitted by law.
Culture & Benefits
Full-time onsite engineering role with other in-office perks.
Stock options and flexible PTO.
100% paid medical, dental, and vision insurance for employees.
Health Savings Account contributions, Flexible Spending Account, and supplementary health benefits.
Paid short- and long-term disability insurance, life and supplemental insurance options, and Employee Assistance Program.
Paid holidays and parental leave.
Будьте осторожны
Если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого. Это мошенники. Обязательно жмите «Пожаловаться» или пишите в поддержку. Подробнее в гайде →