About the Role:
Machine Learning (ML) at The New York Times enhances the experience of our 150 million digital readers from around the globe and grows our subscriber base through content recommendations and personalizations. The Machine Learning & AI team builds and maintains the infrastructure that hosts all of The New York Times real-time ML inference models, including both data and compute. Our partners are Data Scientists that build and deploy their ML models on the ML platform. On the other end, our partners are engineering systems that call these hosted models at scale with low-latency and Service Level Agreements guaranteed by our platform.
As an MLOps Engineer you will partner with product, data science and ML platform engineers to build and maintain the infrastructure that powers the machine learning lifecycle. You will automate and refine the training, deployment, monitoring, and management of our ML models.
This role reports to the Senior Engineering Manager of Data Management Infrastructure.
Responsibilities:
- Build and Automate ML Pipelines: by owning robust CI/CD pipelines for automated model training, validation, deployment, and retraining.
- Productionalize Models: Build the process for packaging, containerizing, and deploying ML models as scalable, low-latency, and highly-available services.
- Monitoring and Operations: Implement and manage comprehensive monitoring for production models, tracking system health, data drift, and model performance degradation.
- Tooling and Infrastructure: Manage and evolve our MLOps toolchain, including model registries, feature stores, experiment tracking systems, and model serving platforms.
- Collaboration and Support: Partner with data scientists to understand model requirements and optimize them for production. Support software engineers in integrating with ML services.
- Best Practices and Governance: Champion and enforce MLOps best practices for reproducibility, versioning (data, code, model), testing, and governance.
Basic Qualifications:
- 2+ years of software engineering or DevOps experience with a focus on MLOps, automation, and infrastructure
- 2+ years of experience programming in Python or Go
- Experience building and managing CI/CD pipelines (e.g., Github Actions, Jenkins, GitLab CI)
- Hands-on experience with containerization and orchestration (e.g., Docker, Kubernetes)
- Cloud platform experience (AWS, GCP) and familiarity with infrastructure-as-code (e.g., Terraform, CloudFormation)
Preferred Qualifications:
- Experience with MLOps tools (e.g., MLflow, Kubeflow)
- Experience with the machine learning model lifecycle, from experimentation to production
- Experience with data processing frameworks (e.g., Spark, Dask, or Ray)
- Experience with low-latency no-sql datastores (BigTable, Dynamo, etc)
- Familiarity with monitoring and observability stacks (e.g., Prometheus, Grafana, Datadog, or ELK)
- Knowledge of data engineering pipelines and orchestration tools (e.g., Airflow, Prefect)