Job Description & Details
This is a heavy-duty MLOps and Data Science gig deeply rooted in the Oil & Gas sector, where you aren't just building offline Jupyter notebooks, but actually owning the end-to-end production lifecycle of ML systems. If you like the idea of building classical machine learning models and ensuring they don't break in high-stakes production environments, this 12-month contract is right in your wheelhouse.
What You'll Actually Be Doing
Your day-to-day will revolve around building, deploying, hosting, and serving classical machine learning models directly on AWS SageMaker. You will spend a significant chunk of your time setting up telemetry, logging, and monitoring dashboards using Prometheus and Grafana to track model drift and performance metrics. It's less about dreaming up experimental algorithms and much more about operationalizing models reliably so stakeholders in the energy sector can trust the outputs.
The Core Tech Stack
To survive here, you need battle-tested experience with AWS SageMaker and a solid foundation in classical machine learning. The team relies heavily on MLOps pipelines, meaning you need to know how to move an artifact from training to a live production endpoint seamlessly. On top of that, fluency in Prometheus and Grafana is non-negotiable because you are directly accountable for the uptime, telemetry, and performance tracking of deployed models.
Interview Expectations
Expect the hiring manager to grill you on how you handle model drift and performance degradation in production, so be ready to talk through a specific time you troubleshot an asynchronous SageMaker endpoint failure using Prometheus metrics. They'll also want to see how you structure your MLOps pipelines for reproducibility, so don't just talk about the models—be ready to whiteboard your CI/CD and deployment architecture on AWS.
Application Advice
Make sure your resume screams AWS SageMaker and MLOps production experience right at the top, since that's what the automated filters and the engineering leads are scanning for first. Explicitly mention tools like Prometheus and Grafana alongside your model deployment stats, and highlight any past experience where you owned the monitoring and telemetry of machine learning systems from scratch.