ML Ops Engineer: Production ML Platform & Reliability

cmcmarkets; Greater London; On-site; GBP 90,000 - 130,000; Full time

Job summary

CMC Markets in London is seeking an experienced ML Ops Engineer to build and operate platform capabilities that move models from experimentation to production. You will own automation, deployment, observability and controls across the ML lifecycle, partnering with researchers, software engineers, and product teams.

This hands-on role focuses on making ML systems reproducible, scalable, secure and dependable—from packaging and release through serving, monitoring, retraining and incident response.

Qualifications

• 3-7 years' professional experience in MLOps, ML platform engineering, ML infrastructure, backend engineering, DevOps or SRE.
• Strong production Python skills, including clean APIs, testing, performance awareness and maintainable services.
• Experience deploying, serving and operating machine-learning models in production environments.
• Practical understanding of the ML lifecycle, including training, validation, inference, model release, monitoring and retraining.
• Experience designing CI/CD workflows and release processes for ML or other production software systems.
• Hands-on experience with at least one workflow or orchestration system used for ML training, validation or deployment.
• Comfort working with cloud infrastructure, containers, infrastructure as code and service networking.
• Strong understanding of observability, monitoring, alerting, incident response and common failure modes in ML systems.
• Ability to reason about system design, reliability and operational trade-offs-not just individual tools.
• Clear communication skills and the ability to work effectively with research, engineering, platform, security and product teams.

Responsibilities

• Build repeatable workflows for model training, validation, promotion, deployment and retraining.
• Productionise models through packaging, versioning, model registry integration, deployment automation and safe rollback.
• Design CI/CD pipelines for ML systems, including automated testing, validation, release controls and environment promotion.
• Manage experiment tracking, model metadata and reproducibility across research and production.
• Build reusable tooling and platform capabilities that support multiple models and engineering teams.
• Deploy and operate batch and online inference services in containerised cloud environments.
• Define and meet availability, latency, throughput and recovery objectives for ML services.
• Monitor service health, infrastructure, data-quality signals, data drift, prediction drift and model performance decay.
• Establish dashboards, alerting and operational runbooks so failures are detected and resolved quickly.
• Support automated or controlled retraining, model promotion, rollback and model retirement.
• Debug production issues across model, application, infrastructure and critical data-dependency layers.
• Improve system robustness, scalability and cost efficiency through automation, observability and infrastructure as code.
• Write production-grade Python for long-running services, deployment tooling and ML workflows.
• Establish testing, validation, release and incident-management practices for ML systems.
• Collaborate with platform, security and data engineering teams on reliable model inputs, access controls, secrets, resilience and compliance.
• Make explicit trade-offs between research flexibility, delivery speed, operational risk and production stability.

Skills

MLOps

Python

CI/CD

Cloud

Containers

Observability

Model serving

Model registry

Experiment tracking

Tools

Docker

Kubernetes

Job description

CMC Markets in London is seeking an experienced ML Ops Engineer to build and operate platform capabilities that move models from experimentation to production. You will own automation, deployment, observability and controls across the ML lifecycle, partnering with researchers, software engineers, and product teams.

This hands-on role focuses on making ML systems reproducible, scalable, secure and dependable—from packaging and release through serving, monitoring, retraining and incident response.