All projects
Homelab

Self-Hosted LXC Ollama Container

Active

A self-hosted AI inference server running Ollama in an LXC container with GPU passthrough for local LLM capabilities.

ProxmoxLXCOllamaAI/MLGPU Passthrough

Highlights

  • GPU passthrough to LXC for accelerated inference
  • Multiple LLM models available (Llama, Mistral, CodeLlama)
  • API endpoint for integration with other homelab services
  • Private and secure local AI without cloud dependencies

Resume Impact

  • Built self-hosted AI inference platform with GPU passthrough for accelerated local LLM capabilities
  • Configured API endpoints enabling integration with other services and applications
  • Managed multiple LLM model deployments (Llama, Mistral, CodeLlama) for various use cases