What is the best practice to deploy a short living process when you have no machine?

I'm new to devops .

I created a converter and want to deploy it . The converter converts a 3D model from a format to another and then you'll be able to visualize the output file on the platform and then you can download the file if you want to.

when bench-marking the process , i found out that for now it runs up to 1 minute when the files are really big , I'm using azure just moved to aws this week . For now the converter converts one file at a time and uses blender's python library and a c++ library ( when trying to use docker i built these inside the container )

I started by creating a docker container that will read a heavy blob and then output the converted file but figured out that docker is not designed to read local files .

I'm searching for the right model to deploy this . Is docker a good solution ? If not , is there other ways to do this ?