Embedded Systems Engineer | Self-Hosted AI Infrastructure

Embedded systems development and maintenance. Building local AI workflows for code analysis, documentation, and automation. Focus on privacy, cost efficiency, and self-hosted solutions.

• Mac Studio M2 Ultra (64GB unified memory) - Primary LLM inference node
  • Serving: LM Studio, MLX-LM
  • Models: Qwen3-Next 80B, Nemotron 3 Nano 30B, Qwen3 Coder 30B
• Razer Blade 15 (i7-12800H, 64GB DDR5, RTX 3070 Ti) - GPU workloads, development
• Firebat MN56 (Ryzen 7 8745HS, 32GB DDR5) - Helper node
• Beelink SER5-MAX (Ryzen 7 6800U, 32GB DDR5) - Proxmox homelab core

• TP-Link Omada Smart SG2008 managed switch
• AdGuard for local DNS
• WireGuard VPN (VPS gateway) - secure remote access to homelab
• Docker/LXC containerization

Workflows

• Code analysis with large context windows (128k+ tokens)
• RAG pipelines for technical documentation
• Learning agents orchestraition tools like Autogen, Langchain

• Languages: C++, Python, Bash
• AI/ML: LM Studio, Ollama, AnythingLLM, Open Web UI
• Infrastructure: Proxmox, Docker, Linux (Debian, Ubuntu, macOS)
• Embedded: Firmware development, hardware integration
• Workflow: Neovim, tmux, Git, Zsh

• Local AI Infrastructure: Self-hosted LLM serving with privacy-first approach
• Homelab Automation: Orchestration and monitoring for distributed inference

Location: Bydgoszcz, Poland
Languages: Polish (native), English (C1)