ML Infrastructure Scheduler Engineer - Kubernetes & GPUs

Isomorphic Labs; Greater London; Hybrid; GBP 70,000 - 120,000; Full time

Benefits offered by this job

Hybrid working

Job summary

Isomorphic Labs in London, UK, is building an interdisciplinary Scheduling team to orchestrate compute for large ML workloads on Kubernetes. You will optimize compute footprint, enable dynamic batch inference, and support scalable training with robust scheduling policies.

The role sits at the intersection of distributed systems and ML infrastructure, requiring collaboration with ML researchers and the broader Infrastructure team to turn research needs into robust platform capabilities and

Qualifications

• Experience building or operating cluster schedulers or batch systems.
• Deep Kubernetes expertise including custom controllers or operators.
• Cloud infrastructure experience with IaC and cloud providers.
• DevOps/SRE practices applicable to AI accelerator infra.
• Proficiency in Python or Go.

Responsibilities

• Design, operate, and extend ML workload scheduling systems for dynamic batch inference and long-running training.
• Implement scheduling policies to maximize GPU/TPU utilization.
• Collaborate with ML researchers and Infrastructure team to translate workload behavior into platform features.
• Build reliable systems with clear failure semantics and robust runbooks/SLOs.
• Develop custom Kubernetes controllers, operators, and CRDs for ML pipelines.

Skills

Scheduling Frameworks

Kubernetes Expertise

Cloud Infrastructure

DevOps/SRE

Python/Go programming

Tools

Kueue

Volcano

Slurm

Job description

Isomorphic Labs in London, UK, is building an interdisciplinary Scheduling team to orchestrate compute for large ML workloads on Kubernetes. You will optimize compute footprint, enable dynamic batch inference, and support scalable training with robust scheduling policies.

The role sits at the intersection of distributed systems and ML infrastructure, requiring collaboration with ML researchers and the broader Infrastructure team to turn research needs into robust platform capabilities and