Gökhan Çalkap

Uses · October 2026

The setup, honestly.

What I actually run day to day, on one laptop. Most of it works with no internet connection at all.

Hardware

GPU
NVIDIA GeForce RTX 5070 Ti laptop GPU, 12 GB VRAM
CPU
AMD Ryzen 9 9955HX
Memory
32 GB DDR5

Local models

Qwen3.6 35B A3B
My daily model. A mixture-of-experts model with about 3B parameters active per token, which is what makes it practical on a 12 GB laptop GPU. A dense model such as Qwen3.8 27B doesn’t fit this hardware.
Qwen Image 2.1
Image generation and editing — concept renders, and the four views that feed 3D generation.
TRELLIS + Pixal3D
Image-to-3D, mixed in one workflow.
ComfyUI
Image generation pipelines.

CAD & printing

SolidWorks, Fusion 360
CAD design.
Ultimaker Cura
Slicing for 3D printing. I ran two printers of my own and produced a lot of parts with them.

Idea to part, offline

How it fits together.

  1. 01

    Idea

    A sentence, a sketch or a need.

  2. 02

    Spec

    A local LLM turns it into a precise description.

    Qwen3.6 35B A3B
  3. 03

    Image

    Concept renders, edited until they’re right.

    Qwen Image 2.1
  4. 04

    Four views

    Front, side, top and perspective — the technical-drawing habit.

    Qwen Image 2.1
  5. 05

    3D model

    Generated from the views and cleaned up.

    TRELLIS + Pixal3D
  6. 06

    Print

    Sliced and printed: a real part.

    Ultimaker Cura + 3D printer

Contact

Working out where AI fits in your operation?

Full-time, part-time or project-based. Based in İzmir; open to Kocaeli, İstanbul and remote.

gokhan@calkaplabs.com or message me on LinkedIn