Computational Scientist / Scientific Machine Learning

The Fusionality Computational Scientist joins the founding engineering team that Fusionality is assembling in Lausanne to build its core technology platform. Fusionality is building the measurement, simulation and control stack for the fusion energy reactors of the future. It spun out of EPFL’s Swiss Plasma Center and describes itself as backed by top investors, with a technology that enables reliable operation, real-time monitoring and analysis of high-temperature plasmas in fusion devices.

The Fusionality Computational Scientist specialises in scientific machine learning (SciML) and suits someone who writes numerical solvers that are fast, correct and differentiable end to end, and wants to see them run against real plasma data and inside real control systems. Fusionality is looking for a PhD-level or senior computational researcher, and its founders have backgrounds at the intersection of fusion physics, advanced control and AI, including Google DeepMind and EPFL.

Key Responsibilities

• Design, implement and validate differentiable solvers in JAX for plasma physics problems such as free-boundary equilibrium, transport and state estimation
• Make the solvers robust and fast enough for production use through advanced numerical techniques for differential equation time stepping, and through optimisation of GPU and TPU performance
• Use automatic differentiation for model calibration against experimental data, parameter estimation, and controller design and optimisation, working with Fusionality’s data and control engineers
• Own the code, including quality, testing, verification against reference solutions, packaging and documentation

Requirements

Essential:

• Substantial code written in a differentiable framework (Fusionality’s stack is in JAX, and Julia or PyTorch experts are welcome if they are happy to switch)
• Strong background in numerical methods for ODEs and PDEs, including discretisation, stiff and implicit solvers, and nonlinear and sparse linear solvers
• Demonstrated ability to operate at the frontier of SciML, including Universal Differential Equations, Neural ODEs and other emerging hybrid data-driven and mechanistic paradigms
• Evidence of independent work at the highest level, such as a PhD or thesis, published code that is used by others, or an industry project you owned
• Proficiency using agentic coding and other modern AI technologies to accelerate your work, while remaining accountable for the results

Desirable:

• Experience with the JAX ecosystem, including Diffrax, Equinox, Optax and Lineax
• Experience with plasma, fluid, electromagnetic or other multiphysics simulation
• Having led the design of a scientific codebase used by others