> Thousands of companies sent their engineering, operations and other teams to these events. This indicates continued strong interest and energy in getting business value from data.
> DataOps, MLOps and now LLMOps are mainstream and part of keynote discussions. The GenAI boom is bringing into sharp focus the need to quickly operationalize models, so it’s no surprise that all these Ops models tailored for the data world are now part of the mainstage discussions in data and analytics focused conferences.
> I had several discussions on the topic of :
“Who should be responsible for running production? “
Is it the NoOps, you build it you run it, model or is Ops a specialized discipline which is better served by Ops specialists? I heard from several leaders of modern technology teams that they want to make sure that engineers are focused on rapidly building new products and focusing on running their products in production detracts from the main mission of rapidly building new products. The counter argument is that putting developers in charge of production allows them to experience the day-to-day operation of their software and more importantly brings them closer to the customer which is vital for the feedback loop. Amazon’s CTO Dr Werner Vogels has been advocate of this model since 2006, he’s often credited for coining the phrase “you build it, you run it.” Former Netflix architect, Adrian Cockcroft wrote a blog Ops, DevOps and PaaS (NoOps) at Netflix which set off some very passionate debates, just read the comments on his blog. After interacting with dozens of data professionals I think a big portion of companies are leaning toward a model which has shared responsibilities between engineers and operations, with focus on collaboration, automation, and unified tooling.
To read the complete blog, click here.
