Description

You will design and deploy time-series ML models and constrained optimization systems for oil and gas production ramp planning and sequence scheduling. You own the end-to-end pipeline from OT data ingestion to operator-centric UI delivery.

This role is hybrid.

Responsibilities

• Build feature engineering pipelines using DHP/WHP, manifold, and upstream/downstream signals to reconcile telemetry with physical intuition.
• Handle OT historian data (PI) at scale, managing gaps, sensor drift, and event slicing for ramp windows.
• Package, deploy, and monitor Azure ML models, including migrating HPC-trained models to cloud-served artifacts via CI/CD.
• Translate complex model outputs into clear operational instructions, such as ramp steps, stoplights, and countdowns for control rooms.
• Implement robust monitoring for data drift and model performance using Application Insights and Azure Monitor.

Required Skills

• 6+ years of experience in machine learning engineering, specifically with time-series forecasting and sensor-level feature engineering.
• Strong proficiency in Python, including NumPy, Pandas, Scikit-learn, and PyTorch.
• Hands-on experience with Azure ML, including pipelines, endpoints, environments, and registries.
• Proficient in Azure Databricks (PySpark, Delta Lake, DLT) and Azure Data Lake Storage (ADLS).
• Experience with CI/CD for ML using Azure DevOps, including YAML pipelines and automated retraining.
• Containerization skills with Docker and ONNX model packaging.
• Ability to build REST APIs using FastAPI or Flask for model serving.

Preferred Skills

• Domain knowledge in oil & gas, energy, or industrial automation, including exposure to artificial lift systems (ESP/gaslift).
• Experience with physics-based modeling, surrogate modeling, or hybrid ML+physics workflows.
• Knowledge of real-time streaming technologies such as Event Hubs, Kafka, or IoT Hub.

Key Skills

Education

Bachelor's degree