Senior MLOps Engineer – Prompt Reliability

Harbor Softworks

Contract · Remote · Flexible hours

About the role

Harbor Softworks is modernizing its legacy production platform, codenamed "PulseStream", to achieve zero‑downtime deployments. As an MLOps Engineer you will design and implement the CI/CD pipelines and observability stack that enable seamless model updates while the system stays live.

Key responsibilities

• Architect a blue‑green deployment strategy using Argo Rollouts for the PulseStream service.
• Containerize legacy ML inference services with Docker and migrate them to a Kubernetes‑based environment on AWS EKS.
• Build automated testing workflows in GitHub Actions that validate model performance and infrastructure health before each release.
• Implement real‑time monitoring with Prometheus, Grafana, and AWS CloudWatch to detect latency spikes during rollouts.
• Collaborate with data scientists to create Terraform modules that provision model versioning endpoints on demand.
• Document runbooks for rollback procedures and conduct a zero‑downtime rehearsal sprint.

Must-have skills

• Advanced Python scripting for pipeline orchestration.
• Deep experience with Docker and Kubernetes in production.
• Proficiency in AWS services (EKS, S3, CloudWatch).
• Strong DevOps mindset with CI/CD tools (GitHub Actions, Argo).
• Knowledge of blue‑green and canary deployment patterns.

Nice to have

• Familiarity with MLflow or DVC for model tracking.
• Experience with serverless components (AWS Lambda) for auxiliary tasks.