Embracing Multi-Tenancy While Scaling MLOps

Explore the challenges and solutions for implementing multi-tenancy in MLOps systems in this 26-minute conference talk. Discover how organizations can efficiently scale their machine learning operations by sharing resources, maintaining isolation, and ensuring security across multiple teams. Learn about key features such as ML workflow isolation, resource quotas, role-based access control, data isolation, and shared artifact repositories that contribute to a secure and efficient multi-tenant environment. Gain insights into overcoming associated challenges in running MLOps efficiently and cost-effectively. See a demonstration of how to implement these concepts using open-source tools like Flyte, enabling different teams within an organization to operate in isolation while sharing resources effectively.

Embracing Multi-Tenancy While Scaling MLOps - Shivay Lamba & Shivanshu Raj Shrivastava