Orkhan Mammadov

Product Designer · AI tools · Creative software · RemoteSan Francisco / New York / Remote

I design the interfaces people use to steer AI — node editors, agent workflows, and tools that stay powerful without getting opaque.

I design generative media tools and build the prototypes that prove them before there’s a spec. Product designer and creative technologist. 20 years from agency UI → real-time installs → GANs → shipping AI product surfaces (ComfyUI, Genie).

[ Case studies ]
[ Marketing website projects ]
[ Own product ]
[ HOW I WORK ]

Hidden complexity is the enemy

Complexity is not. Dense, expert-driven workflows can stay dense — they have to be legible, not simplified.

Systems thinking, craft first

A design system is the bridge between design and engineering, not just internal consistency. Every component maps to a real one.

I prototype in code

React and TypeScript, for the questions a Figma file cannot answer.

Most useful early

When the problem is ambiguous and the interface does not exist yet.

I raise the quality bar across squads

Patterns, reviews, and the boring edge cases that make AI UIs trustworthy.

[ TOOLKIT ]

I prototype in code when a Figma file cannot answer the question, and build the tool when the tool does not exist yet.

Product design

  • Figma
  • Design systems
  • Web, desktop and mobile UI
  • Flows, states and edge cases
  • Prototyping
  • Research and usability testing

Generative AI

  • ComfyUI
  • Image and video diffusion
  • Custom model training
  • GANs and DeepDream
  • LLM and agent workflows

Code

  • TypeScript
  • JavaScript
  • React
  • Next.js
  • ReactFlow
  • shadcn / Radix

Real-time and 3D

  • TouchDesigner
  • Notch
  • Houdini
  • Three.js
  • WebGL
  • WebGPU
  • p5.js
github.com/orkhanart
[ TWO DISCIPLINES, ONE PRACTICE ]
  1. 2000s

    Browser as canvas

    Self-taught web designer as a kid. HTML, JavaScript and Flash, then UI and UX for agencies from 2008.

    HTML · JavaScript · Flash · UI / UX

  2. 2012

    Creative technology

    Real-time visuals and interactive installations. Touchscreen, AR and VR work for Nike, Adidas and Microsoft.

    TouchDesigner · Notch · Houdini · p5.js

  3. 2016

    AI

    DeepDream and GANs first, then custom models trained on my own datasets. That work went to the Venice Biennale in 2019.

    DeepDream · GANs · Custom datasets

  4. NOW

    One practice

    I design generative media tools and build the prototypes that prove them before there’s a spec: node editors, agent workflows, demos, installations.

    ComfyUI · React · Three.js · WebGPU

[ WORKED AT ]
  • Nous Research
  • ComfyUI
  • Genieart
  • Lunchmeat Studio
  • AprilUV
  • Leo Burnett
  • Studio Recode
  • McCann
hey@orkhan.artLet's work together

I read everything and reply to specifics. Include the problem, the timeline, and what you have already tried.

Open to product design roles. Remote.