AI Factory: Accelerating Solutions with Ray

Explore how Lockheed Martin's AI Factory leverages Ray to accelerate AI solutions in this 28-minute conference talk. Dive into the secure end-to-end modular ecosystem designed for training, deploying, and sustaining artificial intelligence applications. Learn about the centrally hosted cloud and bare metal environments, including edge capabilities, that enable users to scale, train, monitor, and serve intelligent systems. Discover how Ray's distributed computing capabilities enhance machine learning workflows across the AI Factory platform, with examples of large object detection models like YoloV5 being trained and deployed. Gain insights into the integration of the full Ray ecosystem, including Ray Spark, KServe, and Ray AIR. Understand the importance of open-source contributions in ensuring authorized and authenticated network traffic. Explore the key features of AI Factory, such as independent Ray Cluster management and the incorporation of Cloud Native Computing Foundation concepts for enhanced security and reliability. Examine the core principles of API-first design, reusable architectures, and user focus in tool assessments. Benefit from the lessons learned and improvements made based on feedback from Ray Summit 2022, offering valuable insights for the Ray community from the perspective of a large organization.