Generative AI, creative tools - 2023–2025

Imake

A web editor that wraps image generation, style transfer, face swapping and upscaling into one interface - built for the moment right after generation, when a result is close but not quite right.

Role
UX/UI design, AI product integration
Platform
Web application
Scope
AI model integration, UX/UI design, full-stack build
Year
2023–2025
Imake's generation screen: a custom prompt field and three style presets - Realistic, Anime and Glamour
The generation screen. A prompt, three style presets, one button - the parameter panel isn't the first thing shown.

The problem

A generated image is a first draft, not an answer

The brief was to put professional-grade generative AI - GANs, diffusion models, transformer architectures, style transfer, face replacement - in front of people who'd never touched a machine-learning tool, without asking them to learn what any of those words mean.

Getting from a text prompt to a plausible image turned out to be the smaller half of that. The harder problem sits one step later: a result that's close but not right, and an interface that has to let someone fix the jacket without losing the face, the light, or the three other things that already worked.

A generated portrait with AI-detected attribute tags overlaid: Brunette, Brown Eyes, Outdoor, Jacket
A result comes back tagged by attribute - Brunette, Brown Eyes, Outdoor, Jacket - instead of as one block of description.
Style transfer shown as a before/after split down the middle of a portrait photo
Style transfer applied to an existing photo, shown as a before/after split rather than a straight replacement.

Approach

Building for people who don't know what a diffusion model is

I worked alongside AI researchers and computer vision engineers who were integrating the generative, style-transfer and face-replacement models, while I focused on the layer wrapped around them - the part a photographer or a small studio actually touches. Inference ran on cloud GPU infrastructure, with in-house optimisation to keep the compute cost of every generation and every edit under control.

One prompt gets you a draft. Everything after that is editing.

Explore feed of AI-generated images, with Regenerate and Edit actions on a selected image and an NSFW content gate on the grid
The Explore feed - every generated image carries its own Regenerate and Edit actions, plus an NSFW gate on the grid itself.

Editing controls

Backing out of one bad step, not the whole image

Negative Prompt lets someone say what to avoid - "don't make it darker" - instead of only what to add. Mask Select scopes an edit to one region of the photo instead of the whole frame. Image Upscaling raises resolution for print or commercial use without softening detail. A Changing History sits underneath all of it, so a face swap, a style pass and an upscale can each be undone or revisited on their own, without redoing the steps that already landed.

Process

From model research to one editor

  1. Research

    Studied generative model architectures - GANs, diffusion models, transformers - and how people actually work inside creative tools.

  2. Design

    Mapped generation, attribute tagging, style transfer, masking and upscaling into one interface a non-technical creative could use.

  3. Build

    Integrated the AI models with the research team, running inference on cloud GPU infrastructure with cost-control optimisation built in-house.

  4. Ship

    Delivered generation, editing and a persistent change history as one product, not a set of separate tools.

Design decisions

Three choices that mattered more than the model

Outcome

Editing became the point, not generation

What shipped bundles generation, style transfer, face replacement, masked editing and upscaling into one web interface that needs no technical background to use, while still giving someone who wants it fine control over what changes and what doesn't.

Output resolution is high enough for print and commercial use, produced on GPU infrastructure built to run in real time rather than a render queue.