top of page

Image-To-Image

Customizing AI Fashion Models Based On User-Uploaded Images

As a product designer at Huhu AI, an AI virtual try-on company, I researched, designed, and delivered a key product feature leveraging the Flux Realism model, successfully securing a $500K business deal with Cider due to its rapid delivery. During the latter pilot testing stage, I implemented client feedback and drove product improvements, increasing internal image generation efficiency by 30%.

ROLE

Product Designer

Product Researcher

Prompt Engineer

TEAM

CEO, 1 Product Manager (Co-Founder), 1 Front-End Engineer, 1 Back-End Engineer, 1 AIGC Scientist, 

DURATION

Oct 2024 - Present

WHAT I DID

AI Model Research

Prompt Engineering

Prompt Template 

PRD Drafting

Wireframing

Prototyping

Design Review Workshop

Context

CONTEXT

Huhu AI is a platform to create custom AI models and generate try-on images.

There are two parts of the service: creating AI models with user-selected traits (e.g., gender, race, age) and generating try-on images. The company targets fast fashion companies aiming to save costs on traditional fashion photoshoots. 

Huhu 1.png

Service 1: Fully Customizable AI models

Huhu 2.png

Service 2: AI Try-on Images

Problem

PROBLEM

Text prompts are insufficient in creating the AI models our clients envision.

Initially, the AI model creation feature was basic: users could select from predefined attributes via dropdown menus, and the system would generate an image based on those selections. However, this approach didn’t address a crucial need: capturing the client’s envisioned aesthetic, which is often better conveyed through images than text prompts.

Problem.png
Need

NEED

We need a feature for users to upload a reference image and generate customizable outputs that match its vibe.

Group 76825.png
Process

PROCESS

Delivering an MVP playground product to secure potential clients

The first part of the project is to build an intuitive interface to serve as a playground to demonstrate our technology during client pitches. I was the sole researcher and designer, delivering the project end-to-end.

Process.png

Iterate the tool for better automation to support enterprise-scale deals

Moving forward to the pilot testing stage with our clients, I was one of three generating AI model and try-on images. My hands-on research and proactive development of SOPs positioned me as a pseudo-product manager, driving product iterations.

Process 2.png
Impact

IMPACT

The playground demo advanced 3 major clients to pilot testing.

Group 76848.png
Shein-Product-Listing.gif

The iterated features drove internal image generation efficiency by ~30%. 

Impact.png
Product Demo

PLAYGROUND DEMO

Capture the Vibe: AI Fashion Models Perfectly Matched to Your Vision

Upload a Reference Image or Choose from Examples to Start 

Huhu-1.gif

Choosing Level of Similarity To The Reference Image

Huhu-2.gif

Vibe-Matched, Customized AI Models Created In Seconds

Huhu-3.gif
Reflection

DESIGN DECISIONS

Converting complex model capabilities into user-friendly interfaces guided by design principles.

The image-to-image feature, built based on Flux Realism model, began as a Gradio demo by our AIGC scientist. Through daily interactions with the Gradio demo, I made key design decisions to convert my understanding of the model's capabilities into intuitive, user-friendly interfaces.

Group 76859.png

Since the product is going to act as a playground demo for client pitches, I focused on two design principles: intuitive and frictionless. Intuitive means clear, bias-free affordances, while frictionless ensures seamless interactions, aligning expectations with actions.

blocker-free.png

001

Selection of Similarity Levels

First, I selected the appropriate components from the design system and tailored their variables based on the documentation and the specific content of the page.

REFLECTION

Experimentation is Key to Human-Centered AI

huhu 1.png

Joining this project just weeks after starting was an exciting journey. By diving deep into our AI model, I gained a solid understanding of its capabilities and limitations, which became the foundation for designing intuitive interfaces. 

huhu 2.png

Navigating ambiguity while running prompt engineering experiments was both challenging and rewarding. By iteratively testing and refining keywords, I uncovered patterns in the model's behavior, which informed both the design and functionality of the feature. 

Efficiency Comes With Good Communication

Huhu 3.png

A 1.5-week timeline required intense focus and adaptability. I ensured continuous alignment with the team to keep us moving efficiently.

Usability Over MVP Constraints

huhu 4.png

Even in the rush to launch, usability remained a priority from my perspective. By prototyping multiple interface options and hosting a design review, I ensured the solution worked seamlessly for users while finding practical compromises with developers. 

bottom of page