MLOps is NOT Real

Integration

The integration of MLOps into DevOps is reshaping how machine learning models are treated within software development. Luis argues that machine learning models should be considered as integral components of intelligent applications, much like any other software component 1. He believes that the current distinction between MLOps and DevOps creates unnecessary complexity and confusion. Chris Benson echoes this sentiment, expressing relief at the idea of MLOps maturing into a seamless part of DevOps 1.

All of that should just be DevOps, you know, people are building and calling that ML ops as well, I feel like creates a lot of confusion.

--- Luis

Luis further explains that the industry is moving towards automation, where machine learning models can be deployed using existing DevOps flows, such as GitHub actions, to streamline the process 2.

   

Standardization

Standardization within MLOps remains a significant challenge, impacting the deployment and operational efficiency of machine learning models. Chris Benson highlights the need for low-code or no-code solutions to simplify model deployment, especially when targeting diverse edge devices 3. Luis envisions a future where automation allows models to be easily integrated into existing DevOps frameworks, reducing the need for specialized knowledge 4.

I feel like with the right tools, you should be able to get a data scientist to export their model into a well defined container.

--- Luis

This approach could bridge the gap between data scientists and DevOps teams, enabling smoother collaboration and more efficient workflows.