Principal ML Ops Engineer (EMEA Remote)

Overview

Build and operate production-grade ML inference infrastructure for real-time AI applications. The role focuses on scalable model serving, distributed GPU systems, reliability, and cost-efficient production operations.

What you'll do

• Build and operate model-serving infrastructure with vLLM, TGI, Triton, or equivalent frameworks.
• Implement blue/green and canary deployment pipelines for ML models.
• Develop auto-scaling, multi-model serving, and intelligent request-routing systems.
• Optimize GPU utilization, memory efficiency, network throughput, and model storage performance.
• Design observability for latency, throughput, GPU usage, costs, and system health.
• Manage model registries and CI/CD pipelines for automated model deployments.
• Own ML system lifecycles, including production support and on-call responsibilities.

What you'll need

• 4+ years of experience in MLOps, platform engineering, SRE, or similar ML infrastructure roles.
• Hands-on experience with model-serving frameworks such as vLLM, TGI, or Triton.
• Experience operating containerized GPU workloads in production.
• Experience with model registries, experiment tracking, and automated deployment pipelines.
• Proficiency in Python and infrastructure-as-code tools such as Terraform or Helm.
• Strong understanding of distributed systems, performance tuning, and production reliability.
• Fluent English and the ability to work independently in a remote-first environment.

Nice to have

• Experience with Kubeflow, MLflow, or KubeAI.
• Knowledge of GPU scheduling, CUDA/ROCm optimization, or multi-tenant inference.
• Experience optimizing costs across GPU types and inference workloads.
• Experience building production systems from scratch or working in early-stage startups.

Details

• Fully remote within the EMEA timezone.
• Start date: ASAP.
• Location: Ukraine.