MLOps Engineer

Role Overview

As an MLOps Engineer, you will bridge machine learning research with robust, production-grade systems. You’ll own the infrastructure, pipelines, and CI/CD workflows needed to deploy, monitor, and scale ML models across the organization, ensuring reproducible training and production inference that meets latency and reliability SLAs.

Responsibilities

• Design, build, and maintain ML pipelines for automated retraining and batch inference (e.g., Kubeflow, Airflow, Prefect)
• Develop and manage feature stores (e.g., Feast or Tecton) to keep feature engineering consistent between offline training and online serving
• Deploy models as high-throughput, low-latency microservices using NVIDIA Triton, KServe, or FastAPI
• Implement monitoring and alerting for model drift, data quality, and system performance using Prometheus, Grafana, and Evidently AI
• Containerize ML workloads with Docker and orchestrate them on Kubernetes across multi-tenant environments
• Build CI/CD pipelines for automated testing, integration, and deployment (e.g., GitOps, GitHub Actions, GitLab CI)

Requirements

• 3–6 years of experience in DevOps, MLOps, or Software Engineering, with a strong focus on ML infrastructure
• Proficiency in Python and familiarity with shell scripting
• Experience with Infrastructure as Code using Terraform
• Cloud platform familiarity: AWS, GCP, or Azure
• Hands-on experience with Kubernetes and containerized application orchestration at scale
• Familiarity with ML lifecycle tools such as MLflow, Weights & Biases, or SageMaker Pipelines
• Strong software engineering practices: version control, automated testing, and code review

Bonus

• Experience deploying LLMs (e.g., vLLM, Ollama)
• Experience with distributed training frameworks (e.g., Ray, Spark)
• BS/MS in Computer Science or related field