I specialize in architecting high-load systems and deploying localized AI infrastructure. My focus is on building secure, independent, and high-performance environments for Large Language Models (LLMs) in the 2026 landscape.
- Virtualization: VMware ESXi 7.0+, Proxmox.
- Operating Systems: Linux (Ubuntu/Debian), Bash scripting for automation.
- Storage Management: Advanced LVM configuration (on-the-fly scaling).
- Containerization: Docker & Docker Compose.
- AI Infrastructure (LLM Ops): Ollama, DeepSeek-R1 (up to 14b), Llama 3.2, Qwen 2.5.
- Hardware Optimization: CPU Inference optimization on Intel Xeon E5 series.
🚀 Featured Project: ai-infrastructure-lab
A comprehensive documentation of my home-lab setup designed for running private AI models.
- Automation: Bash-based telemetry and resource monitoring.
- Case Study: Solving storage bottlenecks for 14b+ models using LVM.
- Hardware: Optimized for 24-core CPU inference with 64GB RAM.
I advocate for "Private AI" — ensuring data privacy and system autonomy by running powerful models on-premise.
- GitHub: Igorvl