The automotive industry heavily relies on AI platforms for model training and serving. However, the fragmented data tool landscape and legacy platforms hinder the scalability and efficiency of AI platforms. This talk addresses the challenges faced by data platform teams, such as slow data loading using PyTorch and Ray, low GPU utilization, and high costs associated with retrieving data from cloud object storage. In this session, Bin Fan and Tarik Bennett will share their insights and solutions from production experiences, demonstrating how Alluxio helps automotive and smart vehicle companies maximize their AI infrastructure.
- Learn more about new trends and challenges faced by data platform teams in the cloud and generative AI era
- Discover a new stack with a combination of PyTorch/Ray and Alluxio to achieve fast data access, high GPU utilization, and lower cloud costs
- Deep dive into real-world production use cases and example journeys for architecting a modern data platform for model training and serving
&nsbp;
