ML Deployment Engineer

About the Role

Our client, a fast-growing deep technology and AI company based in Munich, Germany, is seeking a Machine Learning (ML) Deployment Engineer to build the critical production systems that transition ML models from experimental stages into highly scalable and dependable real-world services. This role sits at the crossroads of ML Engineering, MLOps, and platform engineering, focused on creating robust tooling and infrastructure that ensures deployments are repeatable, well-monitored, and ready for production deployment.

Key Responsibilities

• Design and implement production deployment pipelines specifically for machine learning models.
• Manage the deployment and operation of model-serving workloads on Kubernetes clusters.
• Create scalable inference services by utilizing KServe technology.
• Package and containerize ML workloads using Docker containers for streamlined deployment.
• Develop deployment automation tools and scripts primarily in Python.
• Handle model version control, artifact management, and deployment workflows with MLflow.
• Construct CI/CD pipelines that automate testing and releasing of ML services.
• Deploy and operate workloads across cloud environments including AWS and/or Google Cloud Platform.
• Design and implement strategies for rollout, rollback, and versioning of ML models in production.
• Enhance system reliability, scalability, and observability for ML deployments.
• Automate the workflow from model approval to production endpoint activation.
• Work in close collaboration with ML Engineers to put new models into production without burdening them with infrastructure management.

Qualifications and Skills

• At least 3 years of professional experience in MLOps, ML Engineering, ML Infrastructure, Platform Engineering, or related fields.
• Proficient in Python programming for automation and tooling.
• Experienced with Kubernetes container orchestration.
• Knowledge of Docker for containerizing applications.
• Familiarity with KServe or other comparable model-serving technologies.
• Hands-on experience managing ML lifecycle using MLflow.
• Proficiency in deploying ML workloads on cloud platforms such as AWS and/or GCP.
• Strong understanding and practical experience with CI/CD pipelines and best practices.
• Comprehensive understanding of production machine learning system challenges and requirements.

Additional Desirable Skills

• Experience with Argo CD and GitOps methodologies.
• Familiarity with Kubeflow orchestration platform.
• Knowledge of NVIDIA Triton Inference Server and Ray Serve.
• Experience with ML frameworks such as PyTorch or TensorFlow.
• Expertise in monitoring tools like Prometheus and OpenTelemetry.
• Infrastructure as code using Terraform.
• Experience with advanced deployment strategies like canary or blue-green deployments.
• Managing GPU-enabled inference workloads.
• Capabilities in model monitoring and drift detection.
• Handling real-time inference API operations.