MLOps Engineer

Role Overview

As an MLOps Engineer, you will bridge the gap between ML development and production operations. You will design, build, and maintain automated infrastructure and pipelines that enable reliable deployment, monitoring, and scaling of machine learning models.

Responsibilities

• Design and implement automated CI/CD pipelines for ML models, including testing, packaging, and production deployment.
• Build and maintain orchestration pipelines using tools such as Kubeflow, Airflow, or Prefect for data prep and model training workflows.
• Develop model monitoring and alerting to track model drift, latency, and resource utilization in real time.
• Optimize model serving infrastructure for high-throughput, low-latency inference using Triton Inference Server, TF Serving, or TorchScript.
• Collaborate on centralized feature store architecture to ensure consistent data definitions and reduce training-serving skew.
• Implement containerized deployments with Docker and Kubernetes.
• Manage infrastructure configuration via IaC using Terraform (or similar tools).

Requirements

• 3–6 years experience in software engineering, DevOps, or MLOps; at least 2 years focused on production ML infrastructure.
• Strong Python proficiency.
• Solid Docker and Kubernetes experience.
• Hands-on cloud infrastructure experience (AWS, GCP, or Azure).
• Experience with Terraform (or equivalent IaC).
• Familiarity with ML pipeline/orchestration tools (Kubeflow, Airflow, MLflow) and model registry systems.
• Strong software engineering fundamentals: CI/CD best practices, git workflows, and system monitoring.

Bonus

• Experience with vector databases (e.g., Pinecone, Milvus).
• Experience with Triton Inference Server.
• Experience managing large-scale LLM deployment pipelines.