GT-Edge-AI-Internal/gt-ai-os-community

Supports pluggable inference backends (NVIDIA NIM, Ollama, Groq, vLLM, SGLang) with GPU-accelerated embedding via BAAI/bge-m3, deployable across Ubuntu x86, NVIDIA DGX ARM, and Apple Silicon with multi-tenant architecture separating control panel, user interface, and resource routing layers. Focuses on text-based document RAG workflows with team-based access controls, offline-capable local model execution, and data privacy through zero-retention inference options.