At IKEA, we are a furniture retailing company with a plethora of different personas that work with Data. An evaluation resulted in us having 4 distinct personas that work with data and algorithms at different levels, each following different methods to productionize their AI/ML models. This presented a very clear and obvious challenge as an engineering area. 

-How do we standardize all model development? 

-How do we handle the complexities surrounding Responsible AI?

- How can we be more cost efficient when developing compute intensive models?

- How do we address the needs of personas that are not code first?

- How do we leverage open-source components together custom-built components to produce a cohesive platform for MLOps?

- How do we establish the desired governance for such vast organization?

The answer lies in establishing a custom developed MLOps platform that leverages Open Source, Public Cloud, Private Cloud, and multiple interaction interfaces that meets the need of our data community. A platform that is both code first and UX driven with a consistent user experience throughout. Come, learn, and take experience from our journey of building a MLOps platform with a platform as product mindset!