Назад

Не получаете ответ?

Telegram-вакансии старше 7 дней могут быть уже неактуальны.

8 дней назад

Software Engineer (Scheduler)

Формат работы
hybrid
Тип работы
fulltime
Грейд
senior
Страна
UK
vacancy_detail.hirify_telegram_tooltipВакансия из Telegram канала -

Мэтч & Сопровод

Покажет вашу совместимость и напишет письмо

Описание вакансии

TL;DR
Software Engineer (Scheduler) (Kubernetes/ML Infrastructure): Designing and operating scheduling systems for complex machine learning workloads on Kubernetes with an accent on topology-sensitive placement, gang scheduling, preemption, and GPU/TPU compute. Focus on building custom Kubernetes controllers and operators, translating ML research requirements into platform capabilities, and ensuring reliable failure semantics, runbooks, and SLOs.

Software Engineer (Scheduler)

Company

Isomorphic Labs

Conditions

6 days ago

Skills

Candidate Availability

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will join a newly formed Scheduling team to own and extend the systems that allocate compute to complex machine learning workloads on Kubernetes. You will design scheduling systems for large-scale, topology-sensitive workloads, drive policies like gang scheduling and preemption, work with ML researchers, and build reliable custom Kubernetes controllers, operators, and CRDs.

Requirements

  • Practical experience building or operating cluster schedulers, batch systems, or queuing frameworks (e.g., Kueue, Volcano, Slurm)
  • Deep hands-on Kubernetes experience including custom controllers, operators, and strong familiarity with the K8s scheduler
  • Strong cloud infrastructure experience: compute provisioning, networking, Infrastructure as Code
  • Solid foundation in DevOps/SRE best practices
  • High proficiency in Python or Go

Responsibilities

  • Design, operate, and extend ML workload scheduling systems for batch inference and long-running training
  • Implement scheduling policies such as gang scheduling, topology-aware placement, and preemption
  • Collaborate with ML researchers and the Infrastructure team to troubleshoot bottlenecks and turn research requirements into platform capabilities
  • Build reliable systems with clear failure semantics, runbooks, and SLOs
  • Build and maintain custom Kubernetes controllers, operators, and CRDs

Benefits

  • Hybrid working model

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 →

Текст вакансии взят без изменений

Источник -