Hi, I'm the developer behind Q-Fit. Today, I want to share the honest story of how this service came to be. It's not some grand startup tale — just a story of one backend developer wrestling with AI to build a side project.
It All Started Simply
When Google Gemini 3.0 Pro was released, I started out of pure curiosity — "Let me see how well this AI can actually build things." I just wanted to test it out and share the results with my coworkers. Then a colleague happened to share a psychology test link with me, and while taking it, a thought popped up: "Could I build something like this myself?"
Once I actually started building, the basic framework came together faster than I expected — the AI wrote code like it was nothing. I became convinced I could polish it into a real service, and driven by the question "Can people actually enjoy something I built while I earn some revenue from it?" — it turned into a full-fledged side project.
Tech Stack: A Completely Different World from My Day Job
My day job is as a Java/Kotlin + Spring Boot backend developer with 5 years of experience. Serverless architecture and frontend development were unfamiliar territory for me. So when choosing the tech stack, my criteria were clear: pick technologies that AI can handle well.
- React + Vite: The frontend framework with the most AI training data available
- Firebase (Firestore, Functions): Minimizing backend costs with serverless
- Cloudflare Pages: Chosen because static page deployment is completely free
- Tailwind CSS: Utility CSS that lets you build decent UIs fast even without strong design skills
These were all technologies I was using for the first time, but with AI guiding me step by step, I managed to deploy without barely reading any official documentation. Of course there were bumps along the way, but I pushed forward with the mindset of "just make it work first, improve later."
Developing with AI: Between Convenience and Frustration
For development, I mainly used Google's Antigravity. It had many advantages: generous quotas, image generation, workflow automation, and more. Repetitive tasks especially — like creating new psychology test data or batch-processing multilingual translations — were incredibly convenient once you established a pattern.
But there were clear downsides too. Gemini 3.0 Pro kept ignoring the project guidelines I had written. It would use hardcoded colors instead of following the style guide, or ignore the predefined component structure. I had to review code and request fixes repeatedly, making quality maintenance above a certain bar the biggest challenge.
A Backend Developer's Frontend Struggles
AI helped me get through the implementation itself, but the real challenge was UI/UX design. Being a backend developer, questions like "How do I make users feel comfortable?" and "Is this button placement intuitive?" constantly haunted me. I'm confident designing APIs, but making screens look beautiful and feel natural for users turned out to be a completely different skill.
I still don't think Q-Fit's design is perfect. But with the mindset of "if I wait until it's perfect, it'll never launch," I chose to release first and improve based on user feedback.
Future Plans
Honestly, there's still a mountain of work to do. Just off the top of my head, here's what I have planned.
- Design improvements: Polish the current UI to be more refined and user-friendly
- Continuous content additions: Steadily adding new psychology tests, mini-games, and blog posts
- Google AdSense approval: The realistic goal is to earn even just a few dollars a month
These goals might seem modest to some, but I think that's exactly the charm of a side project. Building what you want without any pressure, enjoying the process itself. Revenue would be nice if it follows, but even if it doesn't, the experience itself is valuable enough.
Closing Thoughts
Q-Fit didn't start from a grand vision. It began from a spark of curiosity and was built step by step with AI as a reliable assistant. There's plenty of room for improvement and a long road ahead, but if someone has a good time through this service, that alone makes it all worthwhile. I'll keep improving it steadily, so please keep an eye on Q-Fit.