Add example on LLM offline inference

[JasonLo]

This example may be valuable for:

• Generating large volumes of synthetic data with open-source LLMs.
• Performing large-scale, structured data extraction from text.
• Embedding extensive text datasets.
• Cost-effectively executing LLM-driven tasks at scale without relying on expensive commercial solutions.

For more details, please refer to vllm_batch_inference/README.md.

[iross]

Thanks for submitting this! Sorry it took so long to look at it 😂. Let me know what you think of the changes I mentioned. I wish the uv fix was less ugly.. Your suggestion of building it from source sounds painful, but at least then we wouldn't be stomping permissions all over /root/...

[agitter]

Thanks for the contribution @JasonLo!

[JasonLo]

Thank you, @iross and @agitter, for your input. I have tested the changes mentioned above, and they seem to be working well now. I tested a job with job_id = 1455169.