On-Premiss Private AI

On-premiss or on-prem AI using Large-Language Models (LLM) are increasingly popular because your inquiries and information stay local and private. Your questions and data stay out of the public domain.

WyoAg.ai typically uses Retrieval-Augmented Generation (RAG) with on-prem AI. A “group of experts” pre-trained model is modified with custom content. We have a demonstration model of on-prem AI with RAG. It is loaded with resources including the Oklahoma State University: Nutrient Requirements of Beef Cattle (E-974), OSU Extension Beef Cattle Nutrition Workbook , Cattle related Merck Veterinary Manual online sections, University of Wyoming Extension Cattle Management Manual (MP-97), NRCS Wyoming / National resources — Grazing Management (Code 528) standards, BQA (Beef Quality Assurance) National Manual, Field Guide, and Cattle Care & Handling Guidelines. Additional documents, spreadsheets, pictures, etc. can be loaded to further customize your on-prem model.

On-prem private AI models have increased in capability and efficiency while at the same time the hardware needed has both decreased in cost and significantly increased in capability.

We will set up a custom model and hardware according to your needs. If you like the results, you can purchase commercial-off-the-shelf hardware for your on-prem AI. The hardware used is typically an Apple mini and kept off the internet. We can also setup Apple Studio and Nvidia DGX systems for advanced requirements.