Technology - ML Ops Engineer

Job summary

The ML Ops Engineer will drive the operation of production-grade Machine Learning and LLM services on Azure, ensuring models run as reliable, scalable, and high-performing systems. This role involves owning the end-to-end MLOps/LLMOps lifecycle, leading on CI/CD, deployment automation, monitoring, and incident response. Working closely with Data Science, the engineer will turn models into robust production services with strong governance, observability, and continuous optimisation.

Benefits

Qualifications

• Strong Python engineering skills
• Experience in ML frameworks such as scikit-learn, PyTorch, or TensorFlow
• Familiarity with experiment tracking
• Comfortable working in regulated environments
• Understanding of privacy, auditability, change control, and handling sensitive data
• Strong DevOps/SRE background, including CI/CD, Infrastructure as Code, monitoring and alerting, incident management, and reliability engineering
• Hands-on experience with containerisation using tools such as Docker and Kubernetes (e.g., AKS)
• Experience with debugging, performance tuning, and working with container registries
• Experience working with Azure, including Azure Machine Learning (pipelines, registries, online and batch endpoints) and Azure Monitor or Log Analytics
• Experience operationalising ML pipelines, including training, batch scoring, feature engineering workflows, and preventing training-serving skew
• Experience implementing safe deployment practices such as blue/green or canary releases, supported by automated validation
• Understanding of data contracts, schema evolution, and data quality practices
• Ability to troubleshoot data drift and missing features

Responsibilities

• Drive the operation of production-grade Machine Learning and LLM services on Azure
• Ensure models run as reliable, scalable, and high-performing systems
• Own the end-to-end MLOps/LLMOps lifecycle
• Lead on CI/CD, deployment automation, monitoring, and incident response
• Turn models into robust production services
• Bring strong governance, observability, and continuous optimisation for fast, safe, and efficient delivery at scale
• Design and operate CI/CD pipelines for ML models and LLM prompt-flows, covering build, test, validation, deployment, and rollback
• Own model registration and promotion across environments, ensuring traceability, governance, and auditability
• Implement safe deployment strategies (e.g., blue/green, canary, champion/challenger)
• Package and deploy containerised inference services and batch pipelines, ensuring repeatability and rapid rollback
• Run ML and LLM services as production-grade systems, defining SLOs/SLIs, dashboards, and alerting
• Lead incident response for runtime issues, including triage, mitigation, recovery, and post-incident reviews
• Develop and maintain operational runbooks covering restart, rollback, secret rotation, and safe-mode scenarios
• Improve service resilience and reduce MTTR through automation (e.g., self-healing, retries, fallbacks, circuit breakers)
• Implement monitoring for availability, latency, errors, resource usage, and job performance
• Monitor data quality including freshness, volume, completeness, schema drift, and distribution changes
• Monitor model performance, including drift and prediction distribution shifts, and track accuracy where labels exist
• Instrument LLM services for token usage, latency, and safety signals, with clear visibility into cost, quotas, and risks
• Manage prompts and workflows as code, including versioning, code reviews, and automated regression testing
• Own production configuration for LLM deployments, including model updates, limits, and safeguards

Skills

Work schedule

Industry

Relocation