About the Role

You will join the Advanced Analytics team of a leading multinational energy company, working on one of the most technically demanding applications of machine learning in the engineering domain. The mission is to augment the upstream design configuration process for complex mechanical equipment - reducing manual iteration cycles before simulation software is ever invoked. This role sits at the intersection of applied ML and mechanical engineering, working directly with senior engineering stakeholders on highly bespoke, domain-specific datasets.

What you Will Do

· Analyze and model FEA-derived engineering datasets - load cases, material properties, geometric configurations - to build ML-based surrogate models that predict optimal design configurations.

· Apply uncertainty quantification techniques to assess model confidence and communicate prediction reliability to engineering stakeholders.

· Implement active learning strategies to intelligently prioritize the most informative simulation runs, minimizing the number of costly iterations required to train robust models.

· Collaborate closely with mechanical engineers and product management to translate domain constraints into well-defined ML problem formulations.

· Select, justify and implement appropriate ML approaches - regression, ensemble methods, neural networks based on dataset characteristics, and defend those choices with technical rigor.

· Build, validate and iterate on models using Python in an AWS environment.

· Work semi-autonomously: take high-level requirements, structure the problem independently and drive execution with minimal oversight.

Must Have

· Mechanical or physical engineering domain experience - hands-on exposure to engineering datasets involving loads, stress, pressure, material properties or structural behavior. Aerospace, subsea, automotive and heavy industry backgrounds are all relevant.

· FEA dataset fluency - you understand what simulation output data looks like, how it is structured and what the physical quantities represent. Operational knowledge of ANSYS or Abaqus is not required, but you must have worked with data generated by tools of this kind.

· Strong ML foundations - deep grounding in mathematics and statistics. You can justify algorithm selection, explain model behavior and diagnose failure modes. You do not treat ML as a black box.

· Uncertainty quantification - applied UQ experience in an engineering or scientific context.

· Active learning - practical experience designing or implementing active learning loops to reduce the number of expensive simulations runs needed to train reliable models.

· Production-quality Python — clean, maintainable ML code built for real engineering workflows.

· Autonomous working style — you can take high-level requirements, structure the problem independently and deliver without daily oversight.

What Will You Set Apart

· Experience with surrogate modelling, design space exploration or simulation data regression.

· Familiarity with FEA/CFD simulation workflows (ANSYS, Abaqus, Fluent) - you do not need to operate these tools, but understanding the workflow is a meaningful advantage.

· Background in physics-informed ML or scientific computing.

· AWS experience for model development and deployment.