A MLOps Platform for All Data Personas

Explore the journey of building a comprehensive MLOps platform at IKEA in this 52-minute conference talk. Discover how the furniture retail giant addressed the challenge of standardizing model development across seven distinct data personas, each with unique methods for productionizing AI/ML models. Learn about the strategies implemented to tackle complexities surrounding Responsible AI, cost efficiency in compute-intensive model development, and meeting the needs of non-code-first personas. Gain insights into leveraging open-source components alongside custom-built solutions to create a cohesive MLOps platform. Understand the approach to establishing governance in a vast organization and the implementation of a platform-as-product mindset. Delve into IKEA's experience of integrating Open Source, Public Cloud, and Private Cloud technologies with multiple interaction interfaces to meet the diverse needs of their data community.

A MLOps Platform for All Data Personas - Karan Honavar, IKEA IT AB