Built and configured a dedicated local AI server for small business use. Assembled the hardware around a Ryzen 9 9700X and dual NVIDIA RTX 5060 Ti GPUs with 64GB RAM. Deployed a full open-source AI stack on Linux including a local LLM, web UI, and task automation layer — no cloud dependency.
Designed and built a dedicated always-on AI workstation from the ground up — both the physical hardware and the full software stack. The goal was to give a small business owner a locally hosted AI assistant capable of handling office productivity tasks, with no reliance on cloud services or subscription APIs.
Assembled the machine from scratch, selecting components specifically for AI inference — high core count, large RAM capacity, and dual GPUs to maximize the VRAM available to the model.
Used to download and run large language models locally. Handles GPU offloading across both cards to fit larger models in VRAM.
Browser-based chat interface deployed via Docker, giving the end user a clean UI to interact with the local LLM and preserving conversation history across sessions.
Self-hosted search engine configured to give the AI the ability to perform web searches without sending queries to external services.
AI agent layer used to configure and extend the assistant's capabilities for task automation and productivity workflows.
Delivered a fully functional, always-on local AI server running open-source models with no cloud dependency. The system is accessible via a web browser on the local network, maintains persistent conversation history, and is configurable for small business automation tasks.