Staff Machine Learning Researcher

Orbital uses AI to build data center hardware that outperforms the competition. Our AI simulates materials at the atomic level, tests millions of hardware configurations in the time traditional methods test hundreds and finds optimal designs - not just good ones. The result is shipped hardware with specs incumbents can't match: 1 MW/rack, PUE < 1.04, 12-week lead times. Every deployment generates field data that improves our AI models. Better models produce better hardware. Better hardware enables more capable AI. The loop is already closed and tightening with each build cycle. We are not just beneficiaries of AI progress, we are a lever on its rate. Data Centers are where we start, because the market is urgent and the specs are demanding. But the AI-accelerated development process we've built - materials discovery, hardware design, manufacturing optimization - applies to any complex physical system. Data Centers are the proof point, not the ceiling. We have sites in London, Canada and the USA, building teams across ML research, Product development, Mechanical engineering, and Chemical engineering. If you want to work where AI meets atoms, we'd like to hear from you. As a Staff Machine Learning Researcher at Orbital, you will architect cutting-edge AI systems for the multi-scale design of physical technologies. When we say multi-scale, we mean it: we build world-class foundation models for simulating both the microscopic motion of atoms and the macroscopic flow of liquids in 1GW data centers. We then co-design across these different scales using the ingenuity of our scientists and engineers, augmented with best-in-class domain agents. In this role you will set exceptionally high technical standards and drive projects from prototype through to production deployment. First and foremost, we want to work with someone with a love of craftsmanship, continual learning, and building systems that scale. We also value low ego, and a genuine passion for using AI to solve major global industrial technology challenges.