Nice tools for MLOps but from a computer vision perspective?

How do you manage the entire machine learning lifecycle for a computer vision application? Most MLOps tools that currently exist seem to be more tailored to tabular data. I am especially interested in deployment and production pipelines that handle the entire process of receiving an image and its metadata, running preprocessing operations and running deep learning inference on the image.

Also, do you use the cloud for production inference? For example, cloud computing services like EC2 and Lambda make it …

application computer computer vision computervision data deployment image lifecycle machine machine learning machine learning lifecycle metadata mlops mlops tools nice perspective pipelines process production running tabular tabular data tools vision