Research Software Engineer

About us Hephaestus Technologies is a London-based startup using machine learning and computational physics to discover new superconducting materials for the wires that will power fusion reactors. Our work sits at the intersection of condensed matter physics, materials simulation and deep learning. We're a small, technical team where your work directly shapes what we build and discover. The role You'll build and improve the simulation and ML tooling at the core of our materials discovery pipeline. You'll work closely with our theorists and experimentalists, turning physical insight into robust, scalable code and models. What you'll do Develop, optimise, and maintain physics simulation code for materials modellingBuild and train deep learning models for materials screeningIntegrate simulation and ML workflows into a reliable, reproducible research pipelineCollaborate with theory and experimental colleagues to prioritise and validate candidate materialsContribute to research direction and help keep our engineering standards highWhat we're looking for A late-stage PhD from a leading university, ideally in physics/chemistry, or equivalent experienceExperience in research software engineering, computational physics or a related roleA demonstrable track record of developing physics simulationsExposure to condensed matter physics and machine-learned interatomic potentialsA solid grasp of deep learning, with hands-on experience building and training models (e.g. PyTorch or JAX)Strong Python skills and good software engineering practiceNice to have Experience with density functional theory (DFT) or related electronic structure methodsKnowledge of superconductivity or superconducting materialsWhat we offer Strong compensation and flexibility to work hybrid or remoteGenerous stock options25 days of annual leaveThe chance to work on a hard, important problem with a direct line to fusion energy