Lead MLOps Engineer — GPU serving & FinOps

Northern Spark Systems is hiring a Lead Lead MLOps Engineer — GPU serving & FinOps on the ML platform team (Hybrid · Manchester, United Kingdom). This is a hands-on role in ML platform & MLOps—not a generic “AI enthusiast” post.

You will own training, deployment, observability, and cost control end to end: problem framing, offline evaluation, safe rollout, and iteration from production signals. Success is measured with clear metrics (quality, latency, cost, or business KPIs)—not slide decks alone.

Northern Spark Systems operates ML infrastructure for tier-1 UK retail and logistics partners. We optimise GPU utilisation, predictable rollouts, and cost visibility so data science teams ship without fire drills.

We are looking for a Lead MLOps Engineer in Manchester (hybrid). You will set direction for inference serving (vLLM, ONNX Runtime, Triton paths) and tighten our FinOps story across regions.

What you will do

• Own Kubernetes-based model serving stacks: autoscaling, quota management, graceful degradation, blue/green and canaries.
• Implement model registry, promotion pipelines, artefact scanning, and integration with CI for safe releases.
• Build dashboards tying GPU hours, inference QPS, and business outcomes; drive cost optimisation initiatives.
• Partner with networking and security for VPC architectures, encryption, and secret management.
• Mentor engineers and negotiate roadmaps with data science stakeholders.

What we look for

• 5+ years in platform or backend roles with measurable ML infra ownership.
• Deep Kubernetes plus cloud (AWS strongly preferred); strong Python and infra-as-code habits.
• Hands-on troubleshooting of GPU workloads, ONNX/vLLM or comparable stacks.
• Experience leading incident response or on-call rotations for inference platforms.
• UK working rights.

Nice to have

• NVIDIA Triton, Ray Serve, or bespoke batch inference schedulers.
• Retail peak-season war stories and capacity planning patterns.