MLOps Engineer

Role Overview

Own the infrastructure, orchestration, and deployment pipelines that move machine learning models from experimental notebooks to resilient, low-latency production services.

Responsibilities

• Architect, build, and maintain ML-focused CI/CD pipelines using tools such as GitHub Actions, Argo Workflows, or Kubeflow Deploy.
• Deploy and operate production ML models, scaling and monitoring with Kubernetes, Docker, and serving engines like NVIDIA Triton or TorchServe.
• Implement automated monitoring for data drift, concept drift, and system performance (latency, throughput).
• Optimize infrastructure cost and utilization for GPU and high-compute workloads via auto-scaling.
• Establish security, reproducibility, and lineage standards for training data, feature stores, and model artifacts.
• Integrate ML models into backend services via gRPC and REST APIs.

Requirements

• 3–6 years experience in MLOps, DevOps, or machine learning engineering focused on production infrastructure.
• Strong proficiency in containerization, Kubernetes, and Infrastructure as Code (Terraform or CloudFormation).
• Hands-on experience with feature stores (Feast, Tecton) and model registries (MLflow, Weights & Biases).
• Strong software engineering fundamentals in Python and Bash for scalable backend services.
• Solid understanding of cloud infrastructure on AWS, GCP, or Azure, including networking, IAM, and GPU management.

Bonus

• Experience deploying and serving large language models (LLMs) using vLLM, TGI, or TensorRT-LLM.