MLOps Engineer — Seattle, WA

About The Role

The role owns the infrastructure that makes ML possible at scale: training pipelines, model serving, CI/CD for ML systems, and the observability layer that catches problems before users do. This is hands-on platform engineering for teams shipping models into production weekly, not maintaining a static stack.

You will work directly with ML engineers and data scientists who depend on fast, reliable iteration loops. The work spans GPU cluster management, distributed training orchestration, low-latency inference serving, and building the guardrails that keep production models healthy.

Key Responsibilities

• Design and maintain CI/CD pipelines for ML workflows using tools like Kubeflow, MLflow, Argo Workflows, or GitHub Actions, enabling one-click model promotion from staging to production

• Build and operate model serving infrastructure with low-latency inference endpoints (KServe, vLLM, Triton, or SageMaker endpoints), including autoscaling, canary deployments, and rollback automation

• Implement monitoring and observability for production ML systems: data drift detection, prediction quality dashboards, latency/SLO alerting using Prometheus, Grafana, and custom pipelines

• Optimize training and inference workloads on Kubernetes clusters with GPUs, including cost optimization, resource quotas, and spot/preemptible instance strategies

• Build feature stores and data versioning systems (Feast, DVC, LakeFS) to ensure reproducibility between training and serving environments

• Harden ML platform security and compliance posture: access controls, secrets management, model artifact lineage, and audit trails

• Partner with ML engineers to profile and debug slow or failing pipelines, then codify fixes into reusable platform tooling

What We Are Looking For

• 3–7 years of experience in MLOps, DevOps, or backend/platform engineering, with at least 2 years supporting ML systems in production

• Strong proficiency in Python and one systems language (Go, Rust, or Java); comfort reading and modifying Terraform/Infrastructure-as-Code

• Deep hands-on experience with Kubernetes, Docker, and cloud platforms (AWS, GCP, or Azure) in production environments

• Experience with ML orchestration and experiment tracking tools: MLflow, Kubeflow, aydkRFF, or Ray

• Solid understanding of distributed training concepts, GPU scheduling, and inference optimization techniques (quantization, batching, model compilation)

• BS in Computer Science, Engineering, or equivalent practical experience

• Bonus: Experience serving LLMs in production (vLLM, TensorRT-LLM, Triton), GPU cost optimization at scale, or contributions to open-source MLOps tooling