I'm a Software Developer and Automation Engineer from Pakistan. I make useful software that helps businesses use modern AI tools better, do less manual work, and save time.

My recent work focuses on applied AI, spatial intelligence, and intelligent automation, especially using open-weight AI models in real-world production environments. I enjoy taking vague or messy requirements and turning them into systems that are simple, reliable, and easy to maintain.

• Applied AI and Computer Vision Solutions
  • Metric 3D Reconstruction: Building end-to-end pipelines that transform 2D images into scale-accurate 3D environments using metric depth estimation and multi-view stereo.
  • Spatial Intelligence: Using Qwen3 VL, Gemma 4, and Llama 3.2 Vision for image understanding, dimension estimation, and zero-shot object detection.
  • Image Generation & Editing: Building ComfyUI workflows with Z Image Turbo, FLUX.1 dev, and Qwen Image Edit for professional asset creation.
  • Local Deployment: Running large-scale models efficiently via llama.cpp for private and cost-effective inference.
• 3D Tooling and Blender Automation
  • Building custom Blender addons and Python scripts to automate digital twin production.
  • Automating geometry optimization, surface retopology, and mesh cleanup for AI-generated 3D models.
  • Procedural generation of complex structures from technical blueprint specifications.
• Automation and RPA
  • Scalable Python-based workflows, from desktop automation to distributed task management.
  • High-performance web scraping and data collection from dynamic sites using Playwright and custom anti-bot handling.
• Full-Stack Development
  • Building unified automation hubs and APIs with FastAPI, WebSockets, and React/Vue.
  • Modernizing legacy systems, such as upgrading production Laravel applications (v6 to v12) and porting frontends to Vite and Vue 3.
• Systems Programming
  • Windows system-level tooling and graphics programming in C++ and Python, including GPU compute with HLSL and CUDA.

• Developing metric-accurate 3D reconstruction tools that bridge the gap between AI generation and real-world photogrammetry.
• Turning AI research into usable features, focusing on making complex 3D and automation workflows accessible to end-users.
• Exploring agentic tooling that combines visual reasoning with automated decision-making.
• Open to freelance, contract, or long-term roles involving AI applications, 3D automation, or data-driven systems.

Almost everything I build is written by hand, line by line, with a lot of care and time put into every project. Over the past few months, I have started integrating AI assistance tools like Claude to improve certain parts of older software or add new features on top of an existing foundation. Whenever that is the case, it will be clearly stated in the repository. I think being upfront about it is part of writing honest, maintainable software.

Website: https://danyalziakhan.com
GitHub: https://github.com/danyalziakhan
LinkedIn: https://www.linkedin.com/in/danyalziakhan
Dev.to: https://dev.to/danyalziakhan
Facebook: https://www.facebook.com/danyal.zia.khan

I like building software that feels useful the moment you start using it. If you're looking for someone who can turn ideas, data, or AI models into working systems, feel free to reach out.