Explore how Apache Beam, an open source tool for building distributed scalable data pipelines, can be used to perform common machine learning tasks, with a focus on running inference at scale. Learn about the challenges of deploying models at scale and gain the ability to use Beam to easily parallelize inference workloads. Watch a demo showcasing how Beam can be used to deploy and update models efficiently on both CPUs and GPUs for inference workloads. Gain a high-level understanding of Beam and its applications in machine learning. The 34-minute talk is presented by Danny McCormick, a committer on the Beam project and a senior software engineer at Google, who brings his expertise in open source communities and experience from working on projects like GitHub Actions.