Omnimage
An AI-powered background generation platform that lets enterprise companies like IKEA, Walmart, and Amazon create hyper-realistic product imagery — without the photoshoot.
Understanding the Problem Space.
Connecting the dots between design knowledge and development — under a real deadline, for a real launch.
Defining the Core Problem
Enterprise companies — IKEA, Walmart, Amazon — spend enormous budgets on product photography. Renting studios, hiring photographers, designing interiors, setting up lighting. The costs compound fast. And the output? Often inconsistent, slow to produce, and hard to scale.
Key Pain Points
- High cost of professional product photography at scale
- Slow turnaround from shoot to publish
- Inconsistency across product lines and markets
- No way to iterate quickly on visual styles
The Opportunity
Use AI to generate hyper-realistic product backgrounds — studio quality, at a fraction of the cost, in minutes instead of days.
Landing page — clean, minimal, value proposition clear from the first second
Shaping the Solution Strategy.
Choosing a grayscale design system, a tight tech stack, and AI-assisted development to move fast without sacrificing quality.
Design Philosophy: Simplicity Wins
After studying AI products extensively, one pattern emerged: grayscale and black-and-white interfaces consistently outperform colorful ones for AI tools. When users are dealing with complex technology, the interface should reduce cognitive load — not add to it.
- Grayscale palette: Black, white, and grays only — letting the product imagery take center stage.
- ShadCN + Next.js + Tailwind: A community-trusted design system customized to feel entirely personal.
- AI-assisted development: Claude Code and Cursor to accelerate without cutting corners
Authentication — split layout, SSO-ready, trust signals from the first interaction
Solving the Hard Problems.
Two design challenges that pushed the boundaries of what I thought I knew — and where discrete mathematics finally made sense.
Challenge One: The Upload Experience
How do you design a place where people upload multiple images, add instructions, write prompts, and understand what's happening — all at once? This wasn't about aesthetics. It was about making complexity feel simple. About making people feel confident with new technology.
Upload flow — AI assistant, instruction review, real-time processing feedback
Challenge Two: The Many-to-Many Problem
Years ago in discrete mathematics at SFU — late nights, 2am study sessions, functions, permutations, combinations — I remember thinking "when will I ever use this?" In this project, it mattered more than I could have imagined. We needed to show four types of image relationships:
Results — master grid, individual output view, version history, reprocessing, and error states
The Full Platform.
From orders management to API docs, integrations, and pricing — every screen designed with one principle: simplicity wins.
Key Takeaways
Design Meets Development
This project proved that a designer who codes can ship faster and with more intentionality. Every pixel had a reason. Every interaction was thought through.
AI as Amplifier
Claude Code and Cursor didn't replace the team — they amplified it. We moved at a pace that wouldn't have been possible otherwise, without sacrificing quality.
Simplicity as Strategy
The grayscale design system wasn't a constraint — it was a deliberate choice. When the technology is complex, the interface must be simple. Users learn faster, trust sooner, and stay longer.
Orders Dashboard
Integrations & API
Settings & Account
Pricing
Three tiers — Starter, Pro, Enterprise. Clear value, no hidden fees.
We're still building. Still learning. Still connecting the dots.
Launch: March 2026