epic: Refactor Inference logic from Server Logic

[dan-jan]

• Dynammic linking of "Inference Engines

[tikikun]

@vansangpfiev to research and estimate feasibility

[vansangpfiev]

I have created new repository for llamacpp backend https://github.com/janhq/cortex.llamacpp. Still need some CI and examples to run cortex.llamacpp standalone.
Basically, cortex_cpp will pull shared library from cortex.llamacpp repository, then load it to use.

[tikikun]

Posted here for more visibility a

The details of the structure of refactoring

[tikikun]

Issues for migration to the new design:

• Cameron:
• feat: Migration from coupled server to engine - Python Sub System #534
• Sang:
• feat: Migration from coupled server to engine - llamacpp #535
• Alan
• feat: Migration from coupled server to engine - trt-llm #536