Dataiku & ModelOps
ModelOps helps deploy Dataiku-created data science algorithms throughout the enterprise.
Cassie Kozyrkov, Chief Decision Scientist at Google, says that to cook great food, you don’t have to know how to build an oven; you just need to know how to USE one. She compares cooking to data science: the magic happens when you use AI, not only when you build algorithms in a tool like Dataiku.
Once you invent algorithms in Dataiku Data Science Studio or DSS, you take AI ingredients from Dakaiku, put them in, push a few buttons, and start USING AI to solve business problems!
Data scientists export their work from DSS into ModelOps via APIs or PMML. Once checked in, any business analyst -- your decision-making chefs -- can search for the ingredients they need. Data engineers carefully connect data to create pipelines. Decision Observers and risk managers review those pipelines for compliance, bias, or risk.
There’s a really important and subtle point here: enterprise data science isn’t a job for just one chef, it’s a Team Sport. Like a great restaurant kitchen, ModelOps helps TEAMS of business analysts, decision-makers, data engineers, AND data scientists work together.
Dataiku is just one appliance in the enterprise analytics toolchain, you’ll also want to use AI in business intelligence tools like Spotfire, Tableau, or PowerBI, any app, any data catalog for smart data classification and matching, or even in real-time automated systems using IoT data.
For example, here’s one-click access to Dataiku models from Spotfire: business analysts simply search approved Dataiku-based pipelines, drag them onto a dashboard, and they’re off.
This is why we call ModelOps the new operating system for AI. We make working with Dataiku as easy as installing apps on your mobile phone or pushing buttons on an oven to make great AI meals.
