MLOps Engineer: Production ML Pipelines & Deployments

Evlo AI; New York (NY); On-site; USD 130,000 - 180,000; Full time

Job summary

Evlo AI in New York is building advanced MLOps infrastructure, owning CI/CD for models, scalable training pipelines, feature stores, and monitoring to catch drift before it breaks.

The MLOps engineer enables data scientists and ML engineers to ship models fast without sacrificing reliability. As model complexity grows and GPU costs matter, the work here directly determines how quickly the entire ML organization can iterate and deploy.

Qualifications

• 3-6 years of experience in MLOps, ML platform engineering, or backend/DevOps with ML infra exposure.
• Strong Python skills plus production experience with Kubernetes, Docker, and at least one orchestrator (Airflow, Kubeflow, Dagster).
• Hands-on experience deploying models in production on AWS, GCP, or Azure with GPU inference.
• Deep understanding of CI/CD for ML: versioning, canaries, shadow traffic, automatic rollback.
• Experience with monitoring/observability tools for ML: drift detection, latency, cost tracking.

Responsibilities

• Build and maintain CI/CD pipelines for ML model deployment using Kubeflow, MLflow, Argo Workflows, or GitHub Actions.
• Design and operate training and inference infrastructure on Kubernetes (EKS/GKE), including GPU scheduling, autoscaling, and cost optimization.
• Implement model serving stacks with TorchServe, NVIDIA Triton, vLLM, or KServe for low-latency, high-throughput needs.
• Establish model monitoring and observability: drift detection, performance metrics, alerting with Evidently, Grafana, Prometheus.
• Develop automated data and feature pipelines using Airflow, dbt, or Spark for reproducibility from raw data to training datasets.
• Manage experiment tracking, model registries, and artifact versioning (MLflow, Weights & Biases) with clear promotion/rollback workflows.
• Write infrastructure-as-code using Terraform and collaborate with data scientists to productionize models without handoffs slowing anyone down.

Skills

Python

Kubernetes

Docker

Airflow

Kubeflow

MLflow

Dagster

GitHub Actions

Education

BS/MS in Computer Science or Engineering

Tools

Kubeflow

MLflow

Argo Workflows

GitHub Actions

TorchServe

NVIDIA Triton

vLLM

KServe

Weights & Biases

Job description

Evlo AI in New York is building advanced MLOps infrastructure, owning CI/CD for models, scalable training pipelines, feature stores, and monitoring to catch drift before it breaks.

The MLOps engineer enables data scientists and ML engineers to ship models fast without sacrificing reliability. As model complexity grows and GPU costs matter, the work here directly determines how quickly the entire ML organization can iterate and deploy.