Machine Learning for Mechanical Engineering

ME 505 treats machine learning as a working tool for mechanical engineering problems rather than as pure CS theory, so the emphasis is on turning physical data — microscopy images, material properties, atomistic simulations — into models that actually predict something useful. You'll spend the semester writing Python, curating real datasets, and building up from linear regression through neural networks to ML potentials for molecular dynamics and reinforcement-learning-driven self-driving labs, capped by a project you present to the class. It's a graduate-level bridge course for ME students who already have the mechanics background but need the data-driven half, and it sets you up for materials informatics research and any modern simulation work where ML surrogates have replaced hand-tuned models.