MLOps Engineer

About the role

You'll lead MLOps engagements: audits of customer ML stacks, reference platform stand-ups, and migrations from ad-hoc notebooks to production-grade pipelines. You'll work alongside customer data scientists who own modeling, you own how the model ships, runs, and stays accountable.

What you'll do

• Stand up reference MLOps platforms: training pipelines, model registry, serving, monitoring, alerts.
• Migrate existing models to the new rails one workstream at a time, with canary rollouts and drift monitoring.
• Build feature stores with online + offline parity; eliminate the silent-drift class of bug.
• Wire drift, latency, and quality monitoring to PagerDuty / Slack so customers learn about regressions before customers do.
• Hand over with runbooks, on-call coverage, and an optional retainer for ongoing operation.

What you bring

• 5+ years deploying ML models to production, not just training them.
• Hands-on with at least one of MLflow / Kubeflow / Metaflow / BentoML / SageMaker / Vertex AI in production.
• Strong Python; comfort with Kubernetes, container builds, and CI/CD pipelines.
• AWS, GCP, or Azure at platform-engineering level: networking, IAM, cost, security.
• Track record of taking a model from notebook to served production endpoint with monitoring.

Nice to have

• Triton / Ray Serve / vLLM / TGI experience.
• Drift-detection tools (Evidently, Arize, WhyLabs).
• Compliance-aware deployments (HIPAA, SOC 2, GDPR).