TR

Dogan Kumova

AI Hardware Engineer

20+ Years in Systems Architecture

If AI is going to really scale — even in extreme environments like space — we have to master the hardware first.

Dogan Kumova
  • Location

    Cupertino, CA

About Me


I build the layer everything else runs on. Twenty-three years of enterprise infrastructure — networks, broadcasts, storage, platforms — since 2023 applied to local LLMs and AI hardware, and now to spacecraft at Cosmic DynamiX.

20+
Years in systems architecture
2023 →
AI hardware & local LLM
2026 →
Cosmic DynamiX

Language

Türkçe · Native

English · Professional working

Read more →

Experience


  1. 2026 →

    AI Hardware Engineer

    Cosmic DynamiX · Cupertino, CA

    Space · AI hardware

    With the core team since the company was founded — evaluating and tuning low-power compute for spacecraft.

  2. 2019 – 2026

    Founder & Consultant

    kumova.net

    AI · LLM

    AI-driven infrastructure: local LLM environments, GPU orchestration, quantized models at the edge.

  3. 2010s

    Infrastructure & CDN Architect

    Billur.TV

    CDN · HA

    High-availability infrastructure and content delivery for large-scale digital platforms.

  4. 2003 – 2019

    IT Specialist

    TÜSİAD, İstanbul

    Enterprise IT

    Sixteen years of mission-critical infrastructure, internet-based live broadcasting and digital transformation.

  5. 2000 – 2003

    Designer & Production Director

    Prestige Publishing · Ulusal Kanal

    Media · Design

    Where the eye came from: print layout, then live television production.

Read more →

Skills & Tools


  • AI hardware evaluation

    Low-power chip architectures, kernel benchmarks, watts per token.

  • Local LLM & MLOps

    Inference rigs, fine-tuning, RAG, vGPU and passthrough orchestration.

  • Infrastructure architecture

    Networks, storage, virtualization — 99.9% availability for critical systems.

  • Broadcast & CDN

    Live streaming systems, VoIP, content delivery at scale.

  • AI hardware
  • LLM infrastructure
  • GPU orchestration (vGPU/Passthrough)
  • Quantization & edge
  • Fine-tuning & RAG
  • VMware vSphere
  • Proxmox
  • Docker
  • Linux / CentOS
  • Windows Server & AD
  • CDN & live streaming
  • VoIP
  • Storage (Synology/QNAP/TrueNAS)
  • Digital publishing (Drupal/WordPress)
Read more →

Selected Work


Private AI on GPU

Cost-efficient private training and deployment on owned hardware.

VGPU · PASSTHROUGH · DATA PRIVACY

Read more →