Job title:

Data Scientist Intern (3 months)

Project: Physics-Informed Machine Learning for Numerical Simulation

About SLB:

We are a global technology company, driving energy innovation for a balanced planet. Together, we create amazing technology that unlocks access to energy for the benefit of all.

At SLB, we recognize that our innovation, creativity, and success stem from our differences. We actively recruit people with a diverse range of backgrounds and cultivate a culture of inclusion that unlocks the benefits of our diversity. We want to ensure that everyone feels a sense of belonging here and we encourage, enable, and empower our people to foster inclusivity, build trust, and demonstrate respect for all across the organization.

Global in outlook, local in practice - and with a united, shared passion for discovering solutions, we hire talented, driven people and support them to succeed, personally and professionally.

Location:

Abingdon, Oxfordshire

Description & Scope

Numerical simulation remains the only reliable method to solve partial differential equations to predict future states of a complex physical system - be it weather, fluid flow, quantum dynamics or orbital mechanics. SLB's state-of-the-art reservoir simulator is used to model such a fluid flow in porous media for various applications, including Carbon Capture and Storage (CCS) and geothermal energy systems. The drawback of traditional numerical methods, however, is that they are computational very intensive and are not practical for many realistic workflows.

In this project, you will work on developing a physics-informed machine learning model to predict how a reservoir system behaves when CO2 is injected into it. Machine Learning models have provably been shown to run orders of magnitude faster than conventional simulators and, once trained, provide a promising alternative or enhancement to traditional solvers. The ultimate goal is to use the developed machine learning model to find optimal locations and volumes of CO2 to inject into a subsurface in order to maximize carbon storage and accelerate the global push towards Net-zero.

Deliverables

As part of the Numerical Simulation team, you will work on developing a physics-informed machine learning model to solve Partial Differential Equations on general grids and geometries. You will have access to high-fidelity 3D simulator data to develop and train novel Neural Operator and Graph Neural Network architectures. You will also be integrating this model into full workflows to show that ML solutions run orders of magnitude faster than traditional methods and will have the opportunity to publish in top-tier ML and Applied Mathematics conferences/journals (ICML, NeurIPs, ICLR etc.)

Required Skills & Qualifications

• PhD in Machine Learning, Applied Mathematics, Physics, Statistics, Computer Science or a related discipline
• Theoretical and practical knowledge in machine learning and deep learning
• Strong programming skills and experience with one of Scikit/PyTorch/Tensorflow etc.
• Ability to quickly understand and implement models from research papers

Experience in one of the following would be advantageous

• Neural Operators, DeepONets or Graph Neural Networks etc
• Knowledge of mass conservation equations and material balance
• Simulation basics (numerical methods for solving PDE systems)

We are open to flexible, hybrid working with a combination of on-site & home working days.

SLB is an equal employment opportunity employer. Qualified applicants are considered without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, or other characteristics protected by law.