MLOps/LLMOps solution on AWS?

Yes, SageMaker and BedRock are services for MLOps and LLMOps, for logging you can use SageMaker experiments, SageMaker Model Monitor for model monitoring, SageMaker model registry for model versioning and BedRock is meant to abstract complexities in SageMaker. SageMaker also integrates with Opensource tools like Mlflow, which you can use as an alternative for model tracking and registry. See these workshops to get started, they cover most of your question https://catalog.workshops.aws/mlops-from-idea-to-production/en-US and LlMops https://catalog.us-east-1.prod.workshops.aws/workshops/958877b7-af54-434e-8133-15bbb7693947/en-US

Hi,

This blog post is very interesting in relation to your question: https://aws.amazon.com/blogs/machine-learning/mlops-foundation-roadmap-for-enterprises-with-amazon-sagemaker/

It describes the full roadmap (incl. the right services to use) for enterprises willing to implement MLOps on AWS.

This one is more specific to LLMOps: https://aws.amazon.com/blogs/machine-learning/fmops-llmops-operationalize-generative-ai-and-differences-with-mlops/

Best,

Didier

We also double clicked on this inquiry on a recent episode of re:Post Live. Be sure to check it out! (Diving Deep into Generative AI Operations)