Ask HN: Using selfhosted LLMs in your company

[pastaheld]

Hi, I am curious about selfhosted / opensource LLMs like Ollama in a non-homelab context.

What kind of applications are you building/integrating with them? Are those mostly experimental or in production already? Are you training models yourself? Are you fine-tuning existing ones? How are you deploying and maintaining the models and surrounding tool stack?

Thanks in advance!