Role overview
We are seeking a skilled Machine Learning Engineer to develop and deploy Graph Neural Network (GNN) based surrogate models that approximate complex physics simulations for oil & gas pipeline and well networks. This is a hands-on role for someone who can build high-fidelity neural network models that replace computationally expensive reservoir and network simulators (Nexus, Prosper).
What you'll work on
Design and implement Neural Network architectures to model flow dynamics in interconnected pipeline networks
Build surrogate models that accurately predict pressure distributions, flow rates, and network behavior under varying operational scenarios (training data is acquired through running simulations of the physics models)
Create data pipelines to extract network topology and simulation results from physics-based models (Nexus/Prosper) and transform them into graph representations
Develop training frameworks that incorporate physics constraints (conservation laws, pressure-flow relationships) into neural network loss functions
Collaborate with petroleum engineers to ensure model predictions align with physical behavior and operational constraints
Implement model monitoring, validation, and continuous improvement workflows
Business trip to Kuwait for first 6-12 months. On-site
What we're looking for
Strong expertise in Graph Neural Networks (GCN, GraphSAGE, Message Passing Networks) with proven implementation experience
Deep understanding of deep learning frameworks (PyTorch Geometric, DGL, or TensorFlow GNN)
Experience building surrogate models or physics-informed neural networks (PINNs) for engineering applications
Proficiency in Python and scientific computing libraries (NumPy, SciPy, Pandas)
Demonstrated ability to work with complex data structures (graphs, time-series, spatial data)
Understanding of optimization techniques and handling large-scale training data
Technical Domain Knowledge:
Understanding of graph theory and network analysis
Experience with data structures and graph representations (adjacency matrices, edge lists, sparse tensors)
Knowledge of hyperparameter tuning, model building and uncertainty quantification in ML models
Ready for a long term business trip to Kuwait for first 6-12 months
Background in petroleum engineering, process engineering, or fluid dynamics
Familiarity with reservoir simulation or pipeline hydraulics
Experience with MLOps practices and model lifecycle management
Publications or open-source contributions in graph ML
Experience deploying ML models in production cloud environments (containerization, API development)
Industry Experience:
Oil & gas industry experience is a strong plus, However, candidates with relevant surrogate modeling experience from other engineering domains encouraged to apply
Educational Background:
MS/PhD in Computer Science, Computational Engineering, Applied Mathematics, or related field preferred
Strong mathematical foundation in linear algebra, graph theory, and numerical methods
Understanding of graph theory and network analysis