As of this past Saturday, I deployed a (fairly beefy) desktop computer in my homelab LAN, which is the “dedicated home” for numerous AI models. I am currently debating between which (among two preferred models) is the best fit for my needs. This being the reason why I am starting a new thread.
When I originally set out to build this AI computer I was interested in playing around with Mistral-7B. After experimenting, Mistral-7B barely utilizes 1/5th of my hardware resources. So I moved up to Mixtral-8x7B and Llama 3. These are the two models I am considering building more complex tools with.
Along these lines, I have access to a working API via Ollama. Meaning my AI computer can send and receive data. More importantly, it can interact with other systems if properly configured. So I would like to begin developing with automation tools such as n8n.
Does anyone have experience building with agentic AI and/or AI-driven automation platforms? If so, are there any recommendations you can provide on how to best begin? How have you integrated AI into your homelab networking?