Machine Learning Engineer – Physics-Based Surrogate Modeling

Confidential Company

Multiple VacanciesEmployer Active

Posted 11 min ago

Experience

3 - 8 Years

Job Location

Al Ahmadi - Kuwait

Education

Master of Technology/Engineering(Computers, Mechanical), Master of Science(Physics), Ph.D/Doctorate(Engineering)

Nationality

Any Nationality

Gender

Male

Vacancy

5 Vacancies

Job Description

Roles & Responsibilities

Build Graph Neural Network (GNN) surrogate models to replace physics simulations for oilfield pipeline & well networks. Contract role (Ahmadi, Kuwait) via Brunel — high impact Digital Oilfield project.


Role overview:

We are hiring a Machine Learning Engineer to develop high-fidelity physics-aware surrogate models for oilfield pipeline and well networks. The role focuses on building Graph Neural Network (GNN) based models that accurately replicate outputs from physics simulators (e.g., Nexus, Prosper) to accelerate decision-making in Digital Oilfield workflows.


Key responsibilities:

• Design and implement GNN architectures (Message Passing Networks, GraphSAGE, etc.) to represent interconnected well/pipeline networks.

• Convert simulation outputs and network topology into graph representations (adjacency lists, sparse tensors).

• Build surrogate and physics-informed models (PINNs) to predict pressures, flow rates and network behaviour under multiple scenarios.

• Integrate physics constraints (mass/energy conservation, flow-pressure relationships) into loss functions and training.

• Work closely with petroleum engineers to validate predictions and align models with operational constraints.

• Develop training/data pipelines, hyperparameter tuning, uncertainty quantification and model validation frameworks.

• Implement model monitoring and CI/CD for model deployments (containerization/API).

• Document models, experiments, and provide handover for production integration.


Desired Candidate Profile

Required skills & experience:

• Proven hands-on experience with Graph Neural Networks (GNN, GraphSAGE, message passing networks).

• Expertise in deep learning frameworks and GNN libraries (PyTorch Geometric, DGL, or TensorFlow GNN).

• Strong Python skills and scientific stack (NumPy, SciPy, Pandas).

• Experience building surrogate models or physics-informed ML (PINNs) for engineering simulations.

• Comfortable with graph data structures (adjacency matrices, edge lists, sparse tensors).

• Good understanding of numerical methods, linear algebra, and uncertainty quantification.

• Ability to collaborate with domain experts (petroleum/process engineers).

Nice to have:

• Background in petroleum engineering, fluid dynamics, or process/pipeline simulation.

• Experience with Nexus/Prosper, Eclipse, or industry simulation tools.

• Exposure to MLOps practices, Docker/Kubernetes, cloud deployments (Azure preferred).

• Publications or open-source contributions in graph ML.

Education:

• MS/PhD in Computer Science, Computational Engineering, Applied Mathematics, or related discipline preferred. Strong Bachelors + experience considered.


Employment Type

    Full Time

Company Industry

  • Oil & Gas
  • Petroleum

Department / Functional Area

  • IT Software

Keywords

  • Graph Neural Networks
  • GNN
  • GraphSAGE
  • Message Passing Network
  • PyTorch Geometric
  • DGL
  • PhysicsInformed Neural Networks
  • PINN
  • Surrogate Modeling
  • Reservoir Simulation
  • Pipeline Simulation
  • Nexus
  • Prosper
  • Python
  • NumPy
  • SciPy
  • Pandas
  • MLOps

Disclaimer: Naukrigulf.com is only a platform to bring jobseekers & employers together. Applicants are advised to research the bonafides of the prospective employer independently. We do NOT endorse any requests for money payments and strictly advice against sharing personal or bank related information. We also recommend you visit Security Advice for more information. If you suspect any fraud or malpractice, email us at abuse@naukrigulf.com

Confidential Company