We are looking to staff
one LiveOps (Machine Learning) Engineer in the Latin America region to support our
Supply Chain Enterprise AI (EAI) use cases. The individual will join an existing team comprising employees and contractors from multiple vendors and will collaborate with users and stakeholders across regions, including the U.S., India, and Malaysia.
Key Responsibilities - Incident management, including triage, diagnosis, resolution, and documentation.
- Solution monitoring and operational support.
- Identifying and implementing improvements, such as code refactoring and automated testing, to enhance maintainability and reliability.
- Tracking and reporting solution health KPIs.
- Maintaining and improving technical and user documentation.
The LiveOps Engineer will be expected to: - Gather, assess, and prioritize issues, collaborating with the LiveOps Lead, Product Owner, and technical teams to implement fixes.
- Proactively monitor model, data, and solution performance, identifying opportunities for improvement.
- Execute continuous improvement initiatives under the guidance of the LiveOps Lead.
- Develop and assetize tools for monitoring, diagnostics, and operational support.
- Maintain, test, and enhance model, data, and CI/CD pipelines.
- Monitor operational metrics and logs to proactively identify, diagnose, and resolve issues.
Required Skills Must-Have - Strong Python programming skills with hands-on experience in debugging, troubleshooting, and code refactoring.
- Strong proficiency in PySpark for large-scale data processing and transformation.
- Experience with Databricks and Azure platforms.
- Experience developing dashboards and operational reporting solutions.
- Experience with data and model monitoring.
- Hands-on experience with CI/CD pipelines using GitHub Actions, including automated build, test, and deployment workflows.
- Strong knowledge of Git version control.
Good-to-Have - Experience with time series forecasting solutions.
- Experience using Kedro to develop structured, modular, and maintainable data pipelines.