AI @ HPC tips

[s-sajid-ali]

We need a rewrite of this whole section and rethink what's needed. I don't think running Shiny on GCP needs to be here.

Current Portal	New Portal	Status	Notes
Large number of small files	Move over to HPC storage
SQLite	Move over to HPC storage guides
Joblib and Dask		Needs porting, but I don't see why they need to be in AI @ HPC
LLM on HPC

[s-sajid-ali]

Add a page similar to https://docs.lxp.lu/howto/llama3-vllm/

Investigate and mention the use of https://github.com/VectorInstitute/vector-inference

[s-sajid-ali]

Integrate the guide on fine tuning LLMs on HPC as part of our docs. Perhaps fine tune a smaller model, like phi4 (https://huggingface.co/microsoft/phi-4).

And if we're feeling proactive, add a guide for LoRA fine tuning on HPC as well!

Fine-tuning-Llama2.pdf

[s-sajid-ali]

Outline added in #54