SQLBits 2023
MLOps and ML-Flops: going from dev to prod
Going from an machine learning model trained on your laptop in a notebook called “trainmodelV1Final_FINAL (1).ipynb” to a system ready to deploy is difficult. However, MLOps (a set of principles to prepare your model for prime time) is here to help! This talk is an introduction to all the elements you need to get your code production-ready - CI/CD, dev/UAT/prod, pipelines, and more! We'll walk through system diagrams, with a focus on Azure, but the takeaways will all be platform agnostic. Make sure your model deployment isn’t an ML-Flop!
Going from an machine learning model trained on your laptop in a notebook called “trainmodelV1Final_FINAL (1).ipynb” to a system ready to deploy is difficult. However, MLOps (a set of principles to prepare your model for prime time) is here to help! This talk is an introduction to all the elements you need to get your code production-ready - CI/CD, dev/UAT/prod, pipelines, and more! We'll walk through system diagrams, with a focus on Azure, but the takeaways will all be platform agnostic. Make sure your model deployment isn’t an ML-Flop!