This AI can think like an engineer—and it just designed a spaceship engine

Lin Kayser, co-founder of Leap 71, is developing a software called Noyron that can autonomously design advanced machinery and products. Noyron is a computational model that integrates physics, manufacturing rules, and engineering logic to create objects that meet specific requirements. Unlike parametric CAD or generative AI, Noyron predicts how an object will perform in real-life conditions and can design new objects from scratch. The system uses machine learning to fine-tune itself with each iteration, making it smarter and more accurate. Leap 71 recently tested Noyron by designing a 5-kilonewton rocket engine, which was successfully built and tested in collaboration with the University of Sheffield and AMCM. The engine was designed entirely by Noyron, and its successful test shows that AI can be a true partner in the engineering process. Noyron operates differently from generative AI solutions, which require human oversight and can produce unpredictable results. Leap 71's software aims to produce end-to-end work, with transparent decision-making and human-machine collaboration. The company believes that Noyron could bring Moore's law to engineering, allowing for rapid iteration and exponential progress in computing capability.