feat: cortex should support `engines init` to reinit engines manually

[louis-jan]

Problem
It might be a good idea to have an additional init command specifically for engine-specific setup

cortex engine tensorrt-llm init

For the edge cases that want to init tensorrt-llm ahead of time, eg docker build

Success Criteria

• Users run cortex engines init tensorrt-llm -> cortex pulls the latest tensorrt-llm engine
• Users run cortex engines init tensorrt-llm@0.1.5 -> cortex pulls exactly the TensorRT-LLM engine version (even pre-release)
• Users run cortex engines remove tensorrt-llm -> cortex removes the engine (all of the files)

Additional context
Add any other context or screenshots about the feature request here.