Machine Learning Engineer – Physics-Based Surrogate Modelling

Are you passionate about combining machine learning and physics to accelerate simulation and optimization in energy systems?
This is an exciting opportunity to join one of our major global clients as a Machine Learning Engineer – Physics-Based Surrogate Modelling, where you will design and deploy Graph Neural Network (GNN) and physics-informed neural network (PINN) models to emulate complex physical simulations for pipeline and well network systems.

The role focuses on developing scalable surrogate models that capture multi-physics flow dynamics, reduce simulation costs, and enhance production optimization workflows. You will work at the intersection of data science, physics, and software engineering, transforming simulation data into actionable machine learning frameworks for predictive performance and real-time decision support.

Key Responsibilities

• Develop, train, and deploy Graph Neural Network (GNN)-based surrogate models to approximate physics-based simulations for pipeline and well networks.
• Design data transformation pipelines to convert numerical simulation outputs into graph representations suitable for deep learning architectures.
• Integrate physics constraints and governing equations into neural network loss functions to ensure physically consistent predictions.
• Implement Physics-Informed Neural Networks (PINNs) and hybrid ML-physics models for flow dynamics and pressure prediction.
• Collaborate with domain experts to interpret simulation data, validate model outputs, and ensure consistency with reservoir and production physics.
• Optimize model training for scalability, accuracy, and generalization across varying simulation conditions.
• Build and automate model evaluation, versioning, and deployment frameworks using MLOps tools.
• Develop visualization and post-processing tools to analyze surrogate model behavior and compare results with high-fidelity simulators.
• Document methodologies, validation reports, and research findings for knowledge transfer and technical review.
• Stay current with emerging trends in scientific machine learning, GNNs, and physics-informed AI frameworks.

Required Qualifications / Experience / Skills

• Bachelor’s or Master’s degree in Computer Science, Applied Mathematics, Petroleum Engineering, Mechanical Engineering, or a related field (Ph.D. preferred).
• Minimum 7+ years of experience in machine learning model development, with focus on physics-based or scientific ML applications.
• Proven expertise in Graph Neural Networks (GNNs) using frameworks such as PyTorch Geometric, DGL, or DeepMind Graph Nets.
• Experience developing Physics-Informed Neural Networks (PINNs) or hybrid ML models incorporating domain constraints.
• Strong proficiency in Python and deep learning frameworks (PyTorch, TensorFlow, or JAX).
• Solid understanding of numerical simulation, PDEs, and fluid dynamics in pipeline or reservoir systems.
• Familiarity with data preprocessing, feature engineering, and large-scale scientific data handling.
• Hands-on experience with MLOps tools (MLflow, Docker, Git, CI/CD pipelines) for model deployment and version control.
• Ability to work cross-functionally with simulation engineers, data scientists, and domain specialists.
• Strong analytical, mathematical modeling, and problem-solving skills with attention to detail.
• Excellent written and verbal communication skills for technical documentation and presentations.

Job Location 100% Remote / Hybrid

Type of Employment Full-time / Permanent

Salary: Negotiable

What You Can Expect from the Employer

• Opportunity to work with a global leader in energy and simulation technology.
• Exposure to cutting-edge AI research and scientific computing.
• Competitive compensation and benefits package.
• Collaborative and innovation-driven environment with professional growth opportunities.

Job Types: Full-time, Permanent

Application Question(s):

• Do you have 7 or more years of experience developing deep learning or physics-based ML models?
• Have you built or deployed Graph Neural Network (GNN) or Physics-Informed Neural Network (PINN) models in production or research settings?
• Are you proficient in Python and experienced with frameworks such as PyTorch Geometric or DGL for physics-based modeling?