About this role
What’s the role? The Multiphysics AI & Product Innovation team develops and applies advanced Scientific Machine Learning, Computational Physics & Chemistry, and AI‑accelerated engineering methods to solve complex & high‑impact industrial challenges across Shell businesses. In this role, you will act as a Scientific Machine Learning specialist and Technical Integrator, developing and deploying physics‑guided AI solutions on high-value, complex Multiphysics problems. Your depth lies in AI algorithmic innovation complemented by a breadth of engineering judgment developed through close collaboration with domain experts and asset teams.Rather than being embedded as a single domain specialist, you will act as a technical integrator: understanding business needs well enough to select, adapt, and design the right scientific ML approaches for bespoke, high‑impact problems. You will engage with a diverse range of Shell businesses, including Low Carbon Fuels, Low Carbon Gas, CCS, and Upstream, working on problems such as Multiphysics asset behaviour and degradation (e.g., corrosion, electrochemical systems), Process and design optimization, model acceleration and decision support, operational monitoring and predictive insights for critical assets.Your value lies in understanding both the business problems worth solving as well as where, why & and under what operational constraints the Scientific ML algorithms work; and translating that understanding into robust, deployable solutions.This is a hands‑on experienced individual contributor role designed for someone who thrives at the intersection of deep science, engineering insight, and AI‑driven acceleration; with a strong commercial mindset and a passion for driving practical, scalable innovation.
What you’ll be doing?
Design and develop Scientific ML and physics‑guided AI methods for Multiphysics and engineering applications, including Physics‑informed and physics‑constrained learning, Hybrid modelling (first‑principles solvers + data‑driven models), Reduced‑order models, neural surrogates, and emulators, Operator learning, graph‑based models, and uncertainty‑aware MLOwning and guiding algorithmic decisions on when to apply PINNs, neural operators, surrogates, classical ML, or simulation‑centric approaches. Work with asset/LOB domain experts to understand asset behaviour, operating constraints, uncertainties, and failure modes. Translate loosely defined engineering challenges into well‑posed Scientific ML problems. Ensure developed models are credible, validated against experimental, simulated, and operational data. Check