MLOps Engineer

Evlo AI; Phoenix (AZ); On-site; USD 130,000 - 210,000; Full time

Job summary

Evlo AI is seeking an experienced ML platform engineer to own the infrastructure that moves models from notebooks into production and keeps them there. You will design CI/CD pipelines for ML workflows, manage GPU workloads, and ensure reproducible deployments with observability and rollback capabilities.

You will collaborate with data scientists, LLM engineers, and SREs to deliver a scalable ML platform used across the organization.

Qualifications

• MLOps, ML engineering, or platform/DevOps with ML workloads.
• Strong Python and Bash skills; debugging distributed systems.
• Production experience with Kubernetes and Docker.
• Hands-on with ML observability or serving stacks.
• IaC and cloud platforms with cost-aware GPU management.

Responsibilities

• Design and maintain CI/CD pipelines for ML workflows with automated testing and model gates.
• Build and operate training and inference infrastructure on Kubernetes with GPU scheduling.
• Implement model serving stacks with versioning, canary deployments, rollback procedures.
• Establish monitoring: latency, throughput, GPU utilization, data drift, alerts.
• Own MLOps tooling: registries, feature stores, experiment tracking, artifact.versioning.
• Build data pipelines with Spark, Airflow, and Kafka; ensure lineage and reproducibility.

Skills

Python

Bash

Distributed systems

Kubernetes

Docker

ML observability

Terraform

Cloud platforms (AWS/GCP/Azure)

Data pipelines (Spark/Airflow/Kafka)

Education

BS in Computer Science or Engineering

Tools

GitHub Actions

GitLab CI

Kubernetes

Docker

Spark

Airflow

Kafka

Terraform

MLflow

KServe

vLLM

Triton

SageMaker

Job description

About The Role

The role owns the infrastructure and automation that gets ML models from notebooks into production - and keeps them there. This means building CI/CD pipelines for models, managing GPU workloads, standing up feature stores and model registries, and making sure every deployment is reproducible, observable, and rollback-ready.

About The Role

The role owns the infrastructure and automation that gets ML models from notebooks into production - and keeps them there. This means building CI/CD pipelines for models, managing GPU workloads, standing up feature stores and model registries, and making sure every deployment is reproducible, observable, and rollback-ready.

The team trains and serves models daily, and inference costs, latency, and reliability all matter. The role sits at the intersection of ML engineering and platform engineering, working directly with data scientists, LLM engineers, and SREs to make the ML platform something the whole org can build on safely.

Key Responsibilities

• Design and maintain CI/CD pipelines for ML workflows using GitHub Actions or GitLab CI, with automated testing, model validation gates, and staged rollouts
• Build and operate training and inference infrastructure on Kubernetes (EKS/GKE), managing GPU scheduling, autoscaling, and cost efficiency
• Implement model serving stacks using vLLM, TorchServe, KServe, or Triton - handling versioning, canary deployments, and rollback procedures
• Establish model monitoring and observability: latency, throughput, GPU utilization, data drift, and prediction quality metrics with automated alerting via Prometheus/Grafana
• Own MLOps tooling end-to-end: model registries (MLflow), feature stores, experiment tracking, and artifact versioning
• Build data pipelines that feed training and inference with Spark, Airflow, and streaming systems (Kafka), ensuring lineage and reproducibility
• Write infrastructure as code with Terraform and contribute to on-call rotation for ML platform reliability

What We Are Looking For

• 3–7 years of combined experience in MLOps, ML engineering, or platform/DevOps engineering with direct ML workloads
• Strong Python and Bash skills; comfort debugging distributed systems across training and serving environments
• Production experience with Kubernetes and Docker, including GPU node pools and workload scheduling
• Hands-on with at least one ML observability or serving stack: MLflow, KServe, vLLM, Triton, or SageMaker endpoints
• Experience with IaC (Terraform) and cloud platforms (AWS, GCP, or Azure), including cost-conscious GPU resource management
• BS in Computer Science, Engineering, or related field, or equivalent practical experience